>Copilot: Do you want me to implement consequences for you or babble on and on about what might entirely be a figment of your imagination (Github is up and you're on a 48 hour bender without sleep)
All these monitoring rules are of the format "when 500 errors > baseline for x minutes". Otherwise you'd have monitoring alerts every second. So it is normal for users to already see errors before github officially counts it as an outage.
Do you know of a single service at a single company that actually does that?
I know all of Gmail, every GCE service I can think of, every AWS service I can think of, Amazon.com, Netflix, and Github all do not page on just a single 500.
I know none of those are particularly "high performance" though. Curious where your experience is coming from.
I've been oncall for a different G service that nearly paged on every error. It used the standard error budget tooling, but on hundreds of user buckets because the engineering around locality-specific configuration was... suspect. Many of these buckets had single-digits user. If a user was on their phone and lost signal, I was paged. Very poor oncall experience.
I worked at a large fintech moving billions of dollars in volume a day.
I had a fairly long tenure, where I maintained multiple key services in critical online payments flow. Authentication, authorization, core business and risk data, as well as some cross-cutting control plane stuff, etc. You needed one or more of our services to take a payment, serve any request from the employee dashboard - pretty much everything hit our services. The entire company ground to a halt without my team.
We paged for every single 500. In instances where a particular class of 500 was spurious or not worth fixing, we would leave it acked or mark it as noise. But typically we'd just put in a fix as soon as possible so we didn't page.
Our graceful shutdown and traffic shaping stack was great, but occasionally we'd get a few pages during deploys or failovers.
Oncall was typically not bad, but when it did get bad it was terrible. I've been involved in huge outages that cost hundreds of millions of dollars. Usually it was the fault of multiple teams having compounding runaway failures rather than one service or bug in particular.
It's inexcusable to have a customer's payments not go through. We engineered around resilience. We had strict five nines SLAs and p99 targets and evaluated our adherence with even the smallest partial outage. Hundreds of other services depended on ours, and downstream impacts were huge, so we had to keep a tight ship.
We didn't have "business hours"-only paging either as our platform was available globally, including a heavy install base in Asia.
Assuming the existence of some kind of network (with zero guarantee of 100% reliability), how does this work in practice? Is each 500 treated as an event that needs investigation, even if the result of that would end up as 'a router dropped something from an internal buffer but the transaction as a whole was re-tried by a parent so the service itself recovered'?
Client network timeout shouldn't result in 500. With 408 and retry you should, dependent on the business criteria, get either an upsert (transaction is retried) or 422 (validation that given entry already exists).
Even if it's "DB in datacenter I tried to save to was hit by meteor" event, you can cater for this not to result in 500 (ie - DB unreachable, retry in a couple of minutes); the question is if you want to.
The sub-service at IBM cloud I worked on had an insanely small error budget such that pages were nearly constant. On call was hell week until a few of us insisted on fixing the issues. The "few" of us were contractors. The employees seemed more than willing to just let the pages continue.
Which makes me think a small amount of random issues which happen even though nothing is broken, is normal everywhere. Especially once move things around on a network, there's potential for a lot more random errors.
Bitflips are something that can happen in consumer-grade RAM, so that tracks (and it's comforting that wayward cosmic rays are a substantial reason for an application's crashes!), but on enterprise servers, they will run ECC RAM that is very resistant to bit flips.
This is why data hoarders who have NASes with lots of space insist on running their servers with ECC RAM despite it being significantly more expensive. Because bit flips, for all intents and purposes, cannot happen. The RAM itself detects and corrects for them.
I wouldn't expect bit flips to be a significant contributor to enterprise problems.
Bitflips specifically may not be; things like network issues, noisy neighbors, row/rack/host maintenance (leading to a downed and migrated host) absolutely are things that happen at high frequency at scale and cause your background level of errors to be more than 0.
It’s where monitoring for 9s is more important at that scale than absolute errors. So long as degradation is graceful or retried it should not be a massive problem.
It does require constant tuning and adjustment though.
that is absolutely not the case for any system of size and scale. that would just burn out the on-call team and not result in improvements. Error rates/budgets are used instead.
It depends what you're monitoring. If it's response codes from user generated queries, then I'd agree with you.
But if it is synthetic queries sent from the monitoring platform, then you control the user agent, payload, and endpoints. So any failed requests are a symptom of a misconfiguration and/or failure that should be investigated. Albeit not necessarily as a P1 priority.
Yeah, no, nobody runs cloud services like that. At AWS most alarms required failures in 3 consecutive 5 minute periods. Critical things could be on 3 consecutive 1 minute windows - but that alarm starts a 15 minute escalation for the oncall engineer to check in, and they have to validate the issue isn't a false alarm before updating the status page would even be considered
Re: "page for all 500s": there's a world of difference between "page me with a critical alert at 3am" and "notify me on Monday morning when my normal workday starts". At the extremes:
If my DB health check endpoint is returning 500s for N consecutive checks over M minutes, yeah, please wake me up at 3am!
If one user hit a weird edge case in form validation and got a one-off 500, please don't! We can fix that on Monday.
Not always easy to distinguish those clearly or configure those business hours rules, but for my team at https://heyoncall.com/ that is the goal -- otherwise your team burns out fast. Waking up someone at 3am has a real cost, so you better be sure it's worth it.
I'm sure you're not in ops. Or in a dev org of a service with decent request rates.
What you're asking for is a service to fail silently. There's no way a service with a decent request rate to have 0 500s. Not when it still sees development.
You only do this when you’re trying to use incident management as a hammer to make a point to somebody whom you have otherwise failed to convince to fix something through persuasive argument. Ie, it’s punitive.
No, monitoring for HTTP response code is a subset of observability and not one that generally gives you the best insights into which subsystems are misbehaving nor why.
There are synthetic tests, where you can generate API request calls or even simulate an entire user journey. These allow you to control the user agent, the payloads, and thus you know anything errors back are actual errors. These are triggered by the observability platform (think like running a cron-job) and thus you're not tied to user activity to see when problems arise.
There are other metrics outside of HTTP response codes too. Think like free RAM, CPU usage, disk space, etc. This is just naming some obvious ones because these types of metrics are generally bespoke to the type of application your monitoring. And with these types of monitors, you'd not just have an alert when things have failed, but ideally have alerts when an irregular trend is showing that things are likely to fail too. This latter type of monitors helps you get ahead of the problem before it become customer facing.
Then you have more traditional stuff like logs. This will also be bespoke to the application. But you'd expect errors in logs to get surfaced quickly. Assuming Github have good hygiene in what's being logged.
Tie that up with APMs, RUM, and other goodies like that and you'll have diagnostics to investigate issues when they appear.
(this is just a super high level view of observability too)
> Even a synthetic probe needs a few failures to trigger an alert.
No doesn’t. That just how most people set it up because it’s an easy sane default.
But if you’re measuring internal APIs then there is a strong argument for alerting early.
> You should not alert on cpu, ram, etc
That’s not true to say as an absolute statement. And a generalisation it heavily depends on the system your monitoring and how it behaves under pressure.
But in any case, I wasn’t suggesting CPU alerts were the end goal. I said:
> these types of metrics are generally bespoke to the type of application your monitoring.
Ie you’ll use metrics but those metrics will be highly specific.
The CPU examples were an illustration as to what a “metric” is (it might seem obvious but not everyone is an expert)
> All these monitoring rules are of the format "when 500 errors > baseline for x minutes". Otherwise you'd have monitoring alerts every second. So it is normal for users to already see errors before github officially counts it as an outage.
Is it true that official service status pages are updated automatically?
Yes, Thais can be be really frustrating when you’re trying to get work done. There needs to be more competition and better alternatives and the LLMs need to offer easier connection to these alternatives.
It's an eye opener. Think about it - today, it was a mistake. But, what if it really happened? What if you really lost access to all your years of hard work? It's a wake up call. A blessing in disguise to store what matters to you the most locally, backed up offline. Never trust any single provider. Be it MS or Google or Apple. RAID is the way.
The last two projects I built I did the CI/CD manually with a small win32 service that polls git and builds+deploys the main service locally. It's barely 200 lines of code. Not much to go wrong. "dotnet publish" is not difficult to wrap.
The latest language models have enabled this sort of thing for me. I can integrate a mini Jenkins into every project within a 5-10 minute prompting session. This sort of code isn't hard. It's just tedious, and the LLMs absolutely rock at boring repetitive stuff. Having a win32 service start up successfully on the very first try is something I haven't experienced until 2026.
That works for relatively simple scenarios. When you have to add deploying sql changes or something having to update something in the cloud, you'd have to include a lot more plumbing.
Hah, I know the feeling. I installed Ubuntu on a PC recently, it obviously happened to be one of the days they got DDOSed and apt repos were unreachable. I had other things to take care of, so I put it aside for the next week or so. It didn't help very much, cause after picking it back up, halfway through, Snapcraft went down.
I vibe coded a script that interacts with both Gitlab and Github via their APIs and I've been using it pretty heavily since this morning. I crossed the streams! Goodness, I didn't know it would be _this_ bad!
Insane, we have to come up with contingency plans now for long-duration GitHub outages because we can't safely do deployments. For a service we're paying thousands of $ per year for even though we host runners ourselves...
Same here. You’d think they could at least separate out the GitHub-hosted and self-hosted runners, so you’re still able to dispatch jobs if the self-hosted runners are down.
I’m not convinced they actually do, because GHE on the cloud tends to have the same problems as the main outages. Probably costs extra to be “single tenant” or whatever
Depending on how many thousands of $ per year, it would probably be cheaper and more reliable to self-host GitLab. It's better in terms of organisational structure (you can have one, including access and secret inheritance), and (personal view) Gitlab-CI is better than GitHub Actions because it doesn't push you towards a JavaScript/NPM style dependency hell. And it's actually fairly easy to self-hosted, with options from a single machine with an omnibus package that handles everything to a full blown autoscaling Kubernetes deployment.
Same thoughts - we use an action to ship to production, its builds an image, pushes it to ECS which triggers a deployment.
We can't be blocked here. Seems silly what we settled on this, but for a long time GitHub had been reliable enough for many years, but things are sliding down the pan as of late.
Maybe we need a split between source management and distribution? The former looks like git[hub] to me, the latter maybe more like a Linux distro repo?
It's always best to be portable - always be able to do builds and releases locally (at least, once you get the keys - it shouldn't be possible by default), then add things like github actions on top as convenience.
It's funny, when we were acquired they started moving us to Github actions but it seems that maybe we should stay on our old crusty self-hosted Jenkins setup...
Is it about funds? Why Github is not catching up with the traffic? I know there's a mass rush on Github recently specially due to Claude Code leading users to use Github. sometimes even persuasive.
It's big enough that every time it goes down, it surely stops somebody from pushing fix for what they currently have broken, so I wonder if status page services see some kind of ripple from github outages.
About an hour ago I was having trouble browsing repo files in the browser and I thought "A disturbance in the force, is Github down?" Refreshed HN and loaded up their status site. Nada.
(Ofc, in a sensible universe, we just brush that off to a JS/Firefox glitch or my ISP.)
And yet, here I am. My code is not compiling, my AI isn't vibing, nonetheless I can't work! Two more hours before I can get off!
I've been against self hosting internal tools for a long time mainly because of the devops and other overhead. But AI based devops makes it so easy now to spin up whatever you want now that I'm reconsidering that. I use a lot of ansible for several of our deployments. At this point, most of that is managed via codex.
For Git, all you technically need is ssh access and some backup strategy for your server. It would be bare bones but workable. And there are of course plenty of OSS things that are a lot nicer than that.
I'm still using gh and gh actions and we are mostly below the freemium layer with that. But it is kind of slow and honestly a dedicated vm plus some high CPU/memory workers we can spin up on a need to have basis might be a lot faster. With GH outages becoming more common, my hand might be forced a bit.
In recent weeks, I've spun up listmonk (mailing list solution), matrix (as a slack alternative), and a few other things specific to our software stack. A github alternative would be more of the same. We don't need a lot.
The main objection is that with more moving parts to worry about, the workload for me also increases. Things need updating, monitoring, backups, alerting (and responding to alerts), etc. That sucks up my time and that is scarce.
Another reason for self hosting these days is that with agentic AI tools, self hosted things are a lot easier to integrate into agentic systems. If it is self hosted, you don't have to worry about API limitations, rate limitations, walled gardens, etc. All the traditional SAAS silos are becoming a problem from that point of view. The more locked down it is, the bigger the motive for moving away from it. That's why we ditched Slack for Matrix. Slack is hopelessly locked down and tedious to deal with. Matrix is super easy for this.
Looks lik a terrible source. Like someone ran Claude on the codebase, didn't analyse the results, then vibe coded a blog post. And the dustri.org link doesn't work for me
It's (a) they're under massively increased load because everyone's vibing up new projects these days, (b) they've been in a weird frankenstein "on azure but also we have our own control plane" state for years and they're pushing to no longer have that be the case.
I don't think vibecoding at Github has much to do with it.
Your speculation is that their competitors would naturally not see a commensurate increase in instability while “only” handling 20% of the same crisis?
I don’t buy the excuse. I want to hitch my wagon to those “mysteriously lucky” competitors. (And have. And haven’t had similar issues to Github, since.)
Competitors would be long tail, so a different mode of traffic entirely. Maybe they get spikes that are more easily whack-a-moled than the constant hammering that GitHub receives.
I think it’s much more than 80%, it’s probably the default recommendation and folks who aren’t technical would just accept it. Probably closer to 95% or more
I started using an agent (Codex) on my repo and it went from a a few dozen clones to thousands (3383 this week). I dunno what the agents are doing to clone the repo so many times -- I'm not running 3000 agents or prompts, maybe 10 or so this week. But if this is typical, a 1000x increase in usage across the board can't be good on the system.
GitHub had a blog post about this recently. They reported a significant uptick in volume (repos created, PRs, etc.), which they attribute to AI usage and tooling.
Microsoft has boasted 30% of their code written by AI.[1] However we could only guess if AI generated code is the issue or something else, or a combination of things.
That being said there was a noticeable trend starting around 2022.[2] That being said they’ve also been doing a big migration to Azure. It’s likely a combination of things.
This gets posted every time GitHub is down. This chart is not accurate. It is based on data scraped from GitHub's status page and that data is missing historical incidents from the pre-Microsoft era.
Yeah, it’s not even consistent with their own incident history. I spot checked it and consistently found incidents with downtime/elevated error rates in months listed as 100.00000% uptime on that chart.
The unofficial and offical charts are both lying. The GitHub one ignores actual outages and the unofficial ones count minor display bugs in minor features as a “github outage”.
It could be many things. Microsoft mismanaging stuff. Azure. Vibe-coded Github. So much AI slop being committed it adds an extra burden on the servers, etc.
Whilst you're waiting for it to come back, try out AGENT-CI (which is a project I built.), which runs GitHub Actions on your machine: https://agent-ci.dev. (Open source, etc.)
No, it's not like "act," because it uses the standard Github runner, the difference is that the control plane is an emulation of api.github.com, because of this we can do all kinds of nice things:
Caching in ~0 ms. Pause on failure, so you can let your AI agent fix it and retry without pushing.
I had extremely bad experience trying to setup act on my Macbook. If this is something that actually works (and doesn't steal my credentials), I'm willing to try it despite AI non-features.
Yea, I've had only barely-success on only a few projects with act. Usually due to steps/scripts that use github-internal APIs, but afaict far from always.
I like that it exists, but what a freaking mess that it's necessary and so difficult to do.
I did not say that, what I said was: It's not like `act` because it's not a rewrite of the runner. It's the standard runner... So the one that actually runs GitHub Actions.
I have tried to use act many times, and many times I've failed.
P.S. pause on failure is also helpful for humans, but I'm trying to be realistic about where the future of programming is going...
What problem is github solving that has led it to become critical infrastructure for so many? Is it that everyone is remote and VPNs are too much of a hassle to give everyone access to a build server? Is the serving as the authoritative auth for development services? Does it provide better compliance reporting? It just isn't apparent to me what github offers that you can't get elsewhere with at the same cost and effort. I've been in some pretty large orgs with distributed personnel, but this just hasn't ever been a problem.
GitHub solved the original "code collaboration" problem, and now it's a default easy way to outsource repo management. It also has the most integrations. A lot of companies grew up using GitHub.
GitHub was, once upon a time, quite stable. Things have changed: more features, more usage, and automated agents.
I know what it does, but why is it such a problem that Actions is down? I think you did kind of answer it: "A lot of companies grew up using GitHub," i.e., they are using it as infrastructure by default, not because it does something that otherwise can't be done.
It’s well integrated into massively underpriced agentic coding (and noncoding) workflows, I doubt there’s much more reason than that. The hip thing to do now is hold all your docs in github instead of notion so your agent can traverse them locally
I initially thought it was because I ran out of action minute, and was about to upgrade my plan
Lucky I came here before hitting the confirm payment button
This is your periodic reminder that Github is growing at ~14x (1400%!) annually. This would be incredible growth for a young, unprofitable, VC-funded startup, even Uber never achieved more than ~3x AFAIK. For a widely-established company that was already very well known and a market leader in its niche for many years? Absolutely unprecedented.
This is a conservative estimate assuming linear growth, the actual number is likely going to be higher. Much higher.
It's not too hard to grow 14X YoY if you start from a hundred customers. If you have hundreds of millions? Yeah, not so easy.
Are there any GitHub Actions-compatible CI services out there that don't rely on their infrastructure? I know of depot's but no others; are these resilient to these outages or do they still lose functionality? I imagine the latter but I don't know.
github actions themselves can be self hosted, its quite nice actually to be able to keep your same patterns as cloud hosted actions and with one line change to the yaml have it running on your own hardware. I do this for actions that take 6-7 hours so I am not burning through the 3000 minutes that come free with my account.
Founder of Depot here. To my knowledge, we are the first engine to support different syntaxes in this compatible way via Depot CI [0]. Great time to try it out and let us know your thoughts! We’ve built a lot of cool stuff into it like parallel steps, custom images, and a full CLI/API interface so you can literally everything without going into the web app.
As someone who partially uses depot but was still affected by this github issue, we obviously haven't moved over enough. We use your runners but github is still blocking us.
Hope you don't mind the public ask, it seems useful for others.
If we're using depot runners, and want to use them directly, or move off of github actions being the controller for when things run: what do you suggest?
Yes, triggering Depot CI via the CLI is the sure fire way to avoid all dependencies on GitHub.
We’d need more details around what you’re seeing. It is true that if auth across GitHub is broken than we can’t copy your actions out to be used by Depot CI. However, we have a solution in the works for that as well.
In short, Depot CI, our own engine and control plane is not dependent on upstream actions control plane. But still has to listen for commit events to know if/when to run jobs on things like PRs. This to is being removed in the future.
Are you able to bring your own runners? Our org is heavily invested in self-hosted runners at this point and have gotten a pretty tremendous value from it. I think we'd be wise to get away from GitHub's control plane but keep running jobs in our own infra.
Is there a tier for open source organizations? Do I have to admin any of AWS that runs behind the scenes or can I pay a fixed price to depot and get it to solve everything out of my way?
I used to use Cirrus CI as an alternative to GitHub Actions and am looking for a new alternative. I wonder if Depot could fit in the same way for my needs. I need to run builds and tests in Windows, Linux and macOS.
We currently use external runners (Blacksmith.sh), but that didn't shield us from this as GitHub actions is still the control plane for triggering and monitoring them.
We're now considering Buildkite (apparently they have a GH actions migration tool) or self hosting something (GitLab CI, maybe even Jenkins), as it looks like that would've kept ticking over since we're still seeing webhooks being triggered today during the downtime.
If you would like less dependence on GitHub for issues and PRs, please check out GitSocial, it stores everything in git itself, making them portable and offline-first.
If you want an alternative to GitHub Actions, you could self-host Forgejo Actions, but I'm not that happy with the design.
I much prefer Woodpecker CI, which is an open source fork of Drone.io. It supports multiple Git backends like GitHub, Gitea, Forgejo, Gitlab, Bitbucket. It supports running jobs locally, on Docker, and on Kubernetes. And there's autoscalers built in for AWS, Hetzner, Linode, Vultr, and Scaleway. There's a bunch of 3rd party plugins (https://woodpecker-ci.org/plugins) for custom integrations. The UX is also very simple, with OAuth used not only for authentication/authorization but also setting up & accessing repos. The system architecture is great, with separate components that run stateless connected to a database, and a custom plugin is any program that takes environment variables and does stdio. The config file is a good balance of ugly YAML and convenience syntax like shell-style parameter expansion variables.
It probably takes less than 15 minutes to install, set up, and run WoodpeckerCI for a small team, so it's not a big investment to try out or host. With the autoscaling plugins it lets you scale your workload up to whatever size. Honestly you could run it on a laptop since it's written Go.
Github measures/reports the SLA of the individual services.
The external page linked above goes the other extreme and considers it a bad status whenever any individual service is degraded.
In reality the majority of people only use 3 or 4 of the core services the majority of the time but since there's no "core services" SLA/uptime the usability of github for the majority of people is slightly obfuscated.
Part of it is that it considers downtime in any of the services GitHub provides as GitHub being down. So if GitHub had 100 different services, and only one of them was down at any given time (but at least one was always down), then it would show 0% uptime.
The main operating model with git is going to go back to decentralized. Setting up and managing something like https://forgejo.org/ is a way better experience than constant interruptions by a faulty service that can't meet demand.
The open source contribution model as we once knew it is dead; you're not going to accept patches from random agents. The risk is way too high. And you can see that increasingly "AI Slop" makes it difficult to be a maintainer of any semblance of a popular repo.
So what's the value? A durable place to store work? hah.
Discovery? That part of Github has always been shitty.
So that leaves.. Github Actions? The thing that is down every other day and has been the subject of a few ~rug pulls~/attempted price hikes that are almost surely coming back?
`github-actions[bot]` was disabled for some time, if that's the actor which does the checkout in your setup it could be related. FWIW it's back to working now.
If you don't want to self-host Gitea/Forgejo, I recommend SourceHut for private repos and Codeberg for public ones. Happy to answer any questions you might have for either based on my experience!
free service is down again, let's everyone that use the service for free complain again!!! (sorry for the sarcastic comment but i find it crazy how people feel they are entitled when it's free)
EDIT: sorry i meant this rant at the one complaining for the free service not for the paid customers (which is unacceptable)
We use TeamCity for CI builds, before that Jenkins. Only accessible from the inside of the network.
Even though it's selfhosted and we don't have a dedicated infrastructure team, I don't remember it ever being down in the last 12 years I have been working here.
This is great because I finally set up Actions yesterday for a new project of mine and of course it’s failing today and thinking I screwed up the yaml.
Shout out to all my SF 5am crew checking if their overnight prs passed CI. Real 597 “member of technical staff” energy. I guess we should expect this, it is a Tuesday!
My first time using GH Actions was last week. GH was so flaky that pulling a submodule failed >50% of the time. I had to write a script to retry pulling the submodule in a loop.
I've done some hacky shit in CI scripts, but none made me more mad than that one.
The future of SRE will be the company putting some amount of money on a prediction market against the site going down and you get to take home the winnings as long as the site stays up.
oh man spent so much time trying to debug what's going on. I have a complex setup with GitHub Actions and self hosted runners so I thought it's something broken in my CI setup
I am trying to refrain my "off topic" rants... but such microsoft github abuse is generating so much hate due to their dominant market position, it is hard.
Another outage at GitHub with actions and pages not working thanks to the AI agents Copilot and Tay.ai creating more issues. Last time this happened was 6 days ago. [0]
This time today it was caused by friendly fire by the automatic suspension of the GitHub Actions bot which is now a "Ghost" user. Since there is no CEO of GitHub to contact it we are just going to see more [1] of this again.
You might need to push a critical change soon, but now you cannot. You won't get any of these issues if you self hosted as I said 6 years ago...[2]
We’ve had GitHub actions for long enough, it’s time for GitHub consequences.
>Copilot: Do you want me to implement consequences for you or babble on and on about what might entirely be a figment of your imagination (Github is up and you're on a 48 hour bender without sleep)
i would like to see consequences for "secure sleep" XD.
My action failed with "Unexpected error fetching GitHub release for tag refs/heads/master: HttpError: Sorry. Your account was suspended"
Which certainly made me shit myself, briefly.
A brownout redefined.
Good thing I'm wearing my brown pants today.
ShitHub
https://www.youtube.com/watch?v=LGeOee7x5lY
Same. It's weird how I always find out that GitHub is down before GitHub does. Took 15 minutes before it appeared on githubstatus.com
All these monitoring rules are of the format "when 500 errors > baseline for x minutes". Otherwise you'd have monitoring alerts every second. So it is normal for users to already see errors before github officially counts it as an outage.
In a high performance service with good maintenance and upkeep, you page for all 500s. A noisy pager forces the team to fix the 500s.
Maybe the Github Actions infrastructure isn't run like that.
edit: my oncall rotation notified on all 500s, 24/7, not just rates - https://news.ycombinator.com/item?id=48279262
Do you know of a single service at a single company that actually does that?
I know all of Gmail, every GCE service I can think of, every AWS service I can think of, Amazon.com, Netflix, and Github all do not page on just a single 500.
I know none of those are particularly "high performance" though. Curious where your experience is coming from.
I've been oncall for a different G service that nearly paged on every error. It used the standard error budget tooling, but on hundreds of user buckets because the engineering around locality-specific configuration was... suspect. Many of these buckets had single-digits user. If a user was on their phone and lost signal, I was paged. Very poor oncall experience.
I worked at a large fintech moving billions of dollars in volume a day.
I had a fairly long tenure, where I maintained multiple key services in critical online payments flow. Authentication, authorization, core business and risk data, as well as some cross-cutting control plane stuff, etc. You needed one or more of our services to take a payment, serve any request from the employee dashboard - pretty much everything hit our services. The entire company ground to a halt without my team.
We paged for every single 500. In instances where a particular class of 500 was spurious or not worth fixing, we would leave it acked or mark it as noise. But typically we'd just put in a fix as soon as possible so we didn't page.
Our graceful shutdown and traffic shaping stack was great, but occasionally we'd get a few pages during deploys or failovers.
Oncall was typically not bad, but when it did get bad it was terrible. I've been involved in huge outages that cost hundreds of millions of dollars. Usually it was the fault of multiple teams having compounding runaway failures rather than one service or bug in particular.
It's inexcusable to have a customer's payments not go through. We engineered around resilience. We had strict five nines SLAs and p99 targets and evaluated our adherence with even the smallest partial outage. Hundreds of other services depended on ours, and downstream impacts were huge, so we had to keep a tight ship.
We didn't have "business hours"-only paging either as our platform was available globally, including a heavy install base in Asia.
> We paged for every single 500.
Assuming the existence of some kind of network (with zero guarantee of 100% reliability), how does this work in practice? Is each 500 treated as an event that needs investigation, even if the result of that would end up as 'a router dropped something from an internal buffer but the transaction as a whole was re-tried by a parent so the service itself recovered'?
A reliability engineer from Jane Street gave a great talk about this, five nine’s of correctness in reporting, etc isn’t enough for the SEC.
https://youtu.be/zR9PpXWsKFQ
Client network timeout shouldn't result in 500. With 408 and retry you should, dependent on the business criteria, get either an upsert (transaction is retried) or 422 (validation that given entry already exists).
Even if it's "DB in datacenter I tried to save to was hit by meteor" event, you can cater for this not to result in 500 (ie - DB unreachable, retry in a couple of minutes); the question is if you want to.
The sub-service at IBM cloud I worked on had an insanely small error budget such that pages were nearly constant. On call was hell week until a few of us insisted on fixing the issues. The "few" of us were contractors. The employees seemed more than willing to just let the pages continue.
Im curious about this: because in my experience (working on smaller services though), a small number of errors is always there, as a "baseline".
Recently there was this: https://news.ycombinator.com/item?id=47252971 "10% of Firefox crashes are caused by bitflips"
Which makes me think a small amount of random issues which happen even though nothing is broken, is normal everywhere. Especially once move things around on a network, there's potential for a lot more random errors.
Bitflips are something that can happen in consumer-grade RAM, so that tracks (and it's comforting that wayward cosmic rays are a substantial reason for an application's crashes!), but on enterprise servers, they will run ECC RAM that is very resistant to bit flips.
This is why data hoarders who have NASes with lots of space insist on running their servers with ECC RAM despite it being significantly more expensive. Because bit flips, for all intents and purposes, cannot happen. The RAM itself detects and corrects for them.
I wouldn't expect bit flips to be a significant contributor to enterprise problems.
Bitflips specifically may not be; things like network issues, noisy neighbors, row/rack/host maintenance (leading to a downed and migrated host) absolutely are things that happen at high frequency at scale and cause your background level of errors to be more than 0.
You've completely missed the point - It's not about bitflips it's about errors that are outside the scope of what's fixable.
It’s where monitoring for 9s is more important at that scale than absolute errors. So long as degradation is graceful or retried it should not be a massive problem.
It does require constant tuning and adjustment though.
that is absolutely not the case for any system of size and scale. that would just burn out the on-call team and not result in improvements. Error rates/budgets are used instead.
It depends what you're monitoring. If it's response codes from user generated queries, then I'd agree with you.
But if it is synthetic queries sent from the monitoring platform, then you control the user agent, payload, and endpoints. So any failed requests are a symptom of a misconfiguration and/or failure that should be investigated. Albeit not necessarily as a P1 priority.
forget it, Jake; it’s Azure
Yeah, no, nobody runs cloud services like that. At AWS most alarms required failures in 3 consecutive 5 minute periods. Critical things could be on 3 consecutive 1 minute windows - but that alarm starts a 15 minute escalation for the oncall engineer to check in, and they have to validate the issue isn't a false alarm before updating the status page would even be considered
Re: "page for all 500s": there's a world of difference between "page me with a critical alert at 3am" and "notify me on Monday morning when my normal workday starts". At the extremes:
If my DB health check endpoint is returning 500s for N consecutive checks over M minutes, yeah, please wake me up at 3am!
If one user hit a weird edge case in form validation and got a one-off 500, please don't! We can fix that on Monday.
Not always easy to distinguish those clearly or configure those business hours rules, but for my team at https://heyoncall.com/ that is the goal -- otherwise your team burns out fast. Waking up someone at 3am has a real cost, so you better be sure it's worth it.
Shouldn't Github be large enough to not have anyone on-call, but just rotate the responsible team around the world?
At least when I worked at a Bigcorp a lot of that was being cut to save costs.
> A noisy pager forces the team to fix the 500s.
I'm sure you're not in ops. Or in a dev org of a service with decent request rates.
What you're asking for is a service to fail silently. There's no way a service with a decent request rate to have 0 500s. Not when it still sees development.
A 50 year old bank API? Maybe...
You only do this when you’re trying to use incident management as a hammer to make a point to somebody whom you have otherwise failed to convince to fix something through persuasive argument. Ie, it’s punitive.
You'd expect them to be monitoring more than just the HTTP response codes from user requests for precisely this reason.
If the first they hear of an outage is when user requests start to fail, then that's a failure in their monitoring as well.
But effective monitoring is harder than people assume.
> If the first they hear of an outage is when user requests start to fail, then that's a failure in their monitoring as well.
Isn't that what monitoring actually is? The issue seems to be in their testing, not monitoring.
No, monitoring for HTTP response code is a subset of observability and not one that generally gives you the best insights into which subsystems are misbehaving nor why.
There are synthetic tests, where you can generate API request calls or even simulate an entire user journey. These allow you to control the user agent, the payloads, and thus you know anything errors back are actual errors. These are triggered by the observability platform (think like running a cron-job) and thus you're not tied to user activity to see when problems arise.
There are other metrics outside of HTTP response codes too. Think like free RAM, CPU usage, disk space, etc. This is just naming some obvious ones because these types of metrics are generally bespoke to the type of application your monitoring. And with these types of monitors, you'd not just have an alert when things have failed, but ideally have alerts when an irregular trend is showing that things are likely to fail too. This latter type of monitors helps you get ahead of the problem before it become customer facing.
Then you have more traditional stuff like logs. This will also be bespoke to the application. But you'd expect errors in logs to get surfaced quickly. Assuming Github have good hygiene in what's being logged.
Tie that up with APMs, RUM, and other goodies like that and you'll have diagnostics to investigate issues when they appear.
(this is just a super high level view of observability too)
Even a synthetic probe needs a few failures to trigger an alert.
You should not alert on cpu, ram, etc
> Even a synthetic probe needs a few failures to trigger an alert.
No doesn’t. That just how most people set it up because it’s an easy sane default.
But if you’re measuring internal APIs then there is a strong argument for alerting early.
> You should not alert on cpu, ram, etc
That’s not true to say as an absolute statement. And a generalisation it heavily depends on the system your monitoring and how it behaves under pressure.
But in any case, I wasn’t suggesting CPU alerts were the end goal. I said:
> these types of metrics are generally bespoke to the type of application your monitoring.
Ie you’ll use metrics but those metrics will be highly specific.
The CPU examples were an illustration as to what a “metric” is (it might seem obvious but not everyone is an expert)
> All these monitoring rules are of the format "when 500 errors > baseline for x minutes". Otherwise you'd have monitoring alerts every second. So it is normal for users to already see errors before github officially counts it as an outage.
Is it true that official service status pages are updated automatically?
> it true that official service status pages are updated automatically?
Depends. Typically no because there’s an art to crafting the actual message around impact… but sometimes yes it is automated
More likely that 'update the Status site' lives a long way down their incident response plan, and they have alarms going off well before that
yeah I mean a company the size of GitHub certainly can’t be expected to have enough staff to walk and chew gum at the same time
If it's like other BigTechs I have worked at, you need director-level signoff and comms team approval to post an outage notice
it should be automatic tho. Probably isn't so they can at least get the one nine on availability
Marketing definitely takes interest in status sites
> It's weird how I always find out that GitHub is down before GitHub does
No, it's not. Official updates = potential SLA penalties. Always requires approval.
Yes, Thais can be be really frustrating when you’re trying to get work done. There needs to be more competition and better alternatives and the LLMs need to offer easier connection to these alternatives.
What do the Thai people have to do with this? :(
Reminded me of the "Thai Fighter" joke from family guy's star wars spoof lol
Pretty sure that they wanted to write "this", typed something different by accident, and auto-correct struck.
Oh gee thanks
It's an eye opener. Think about it - today, it was a mistake. But, what if it really happened? What if you really lost access to all your years of hard work? It's a wake up call. A blessing in disguise to store what matters to you the most locally, backed up offline. Never trust any single provider. Be it MS or Google or Apple. RAID is the way.
The last two projects I built I did the CI/CD manually with a small win32 service that polls git and builds+deploys the main service locally. It's barely 200 lines of code. Not much to go wrong. "dotnet publish" is not difficult to wrap.
The latest language models have enabled this sort of thing for me. I can integrate a mini Jenkins into every project within a 5-10 minute prompting session. This sort of code isn't hard. It's just tedious, and the LLMs absolutely rock at boring repetitive stuff. Having a win32 service start up successfully on the very first try is something I haven't experienced until 2026.
That works for relatively simple scenarios. When you have to add deploying sql changes or something having to update something in the cloud, you'd have to include a lot more plumbing.
Deploying SQL changes? Why not just let the application do that on startup. Ofcourse be backward and forward compatible. SQL change only deploy.
"Update something in the cloud" <- What do you mean?
Wasn’t my fault this time! I haven’t started work yet.
https://news.ycombinator.com/item?id=47237377
Yeah but you thought about it, didn’t you?
I did....maybe my powers are growing.
Was about to send my bill to you.
... You're off the hook this time./s
Uh oh. That means there's at least one more like you out there that we don't know about.
I always wanted superpowers, but I never dreamed it'd be like this.
- So many super-heroes/super-villains
Hah, I know the feeling. I installed Ubuntu on a PC recently, it obviously happened to be one of the days they got DDOSed and apt repos were unreachable. I had other things to take care of, so I put it aside for the next week or so. It didn't help very much, cause after picking it back up, halfway through, Snapcraft went down.
Next thing you're gonna tell us you're SRE at GitHub.
Sorry guys it might be me.
I vibe coded a script that interacts with both Gitlab and Github via their APIs and I've been using it pretty heavily since this morning. I crossed the streams! Goodness, I didn't know it would be _this_ bad!
It's only natural that this kind of promiscuity provoked an allergic reaction from Microslop.
Insane, we have to come up with contingency plans now for long-duration GitHub outages because we can't safely do deployments. For a service we're paying thousands of $ per year for even though we host runners ourselves...
Same here. You’d think they could at least separate out the GitHub-hosted and self-hosted runners, so you’re still able to dispatch jobs if the self-hosted runners are down.
If the job queue is down, that wouldn't help, would it?
On my repo the jobs do not get scheduled on the PRs at all, so I assume that separation wouldn't help for todays issue.
They have the github enterprise domain separated out and its working fine right now https://us.githubstatus.com/posts/dashboard
I’m not convinced they actually do, because GHE on the cloud tends to have the same problems as the main outages. Probably costs extra to be “single tenant” or whatever
Depending on how many thousands of $ per year, it would probably be cheaper and more reliable to self-host GitLab. It's better in terms of organisational structure (you can have one, including access and secret inheritance), and (personal view) Gitlab-CI is better than GitHub Actions because it doesn't push you towards a JavaScript/NPM style dependency hell. And it's actually fairly easy to self-hosted, with options from a single machine with an omnibus package that handles everything to a full blown autoscaling Kubernetes deployment.
Sounds good until you see their cvedetails page
I mean, the GitHub Actions supply chain risks and attacks definitely compensate for any GitLab security vulnerabilities you can think of.
Hide it behind VPN, so it's not accessible from outside.
When you own it you can just limit it into vpn-ed company users, that significantly cuts down on the area that can be hit
Same thoughts - we use an action to ship to production, its builds an image, pushes it to ECS which triggers a deployment.
We can't be blocked here. Seems silly what we settled on this, but for a long time GitHub had been reliable enough for many years, but things are sliding down the pan as of late.
Sounds like a very easy process to rewrite in bash/python and have it on hand if needed.
It is a control pain
> For a service we're paying thousands of $ per year for even though we host runners ourselves...
Wait until you charge you for self-hosting runners.
Oh wait. They already tried.
./deploy.sh
Sure. Don't use GitHub.
You can now hire me as an overpriced consultant instead of paying Microsoft.
You should never entirely depend on a third party service for deployments.
Been burned too many times on that one.
Ok.
Move to EC2.
Darn AWS is down.
Alright, run it on a Mac Mini in your basement. Ahh dawn, your ISP is having issues. Good thing you have a backup 5G hotspot.
Ohh no, the power is out.
Eventually you have to trust someone else.
GitHub is a tragedy of the Commons. Too many people are using it, and Microsoft isn't willing to handle it correctly.
Feels like a very good business opportunity. Minimum 50k yearly contracts, GitHub with actual uptime. GitPro ?
Maybe we need a split between source management and distribution? The former looks like git[hub] to me, the latter maybe more like a Linux distro repo?
It's always best to be portable - always be able to do builds and releases locally (at least, once you get the keys - it shouldn't be possible by default), then add things like github actions on top as convenience.
Self host gitlab. If you already host runners it’s not a big lift.
It's funny, when we were acquired they started moving us to Github actions but it seems that maybe we should stay on our old crusty self-hosted Jenkins setup...
Is it about funds? Why Github is not catching up with the traffic? I know there's a mass rush on Github recently specially due to Claude Code leading users to use Github. sometimes even persuasive.
Because scaling complex systems is not trivial
It was pretty easy before October 2018, when Microsoft bought them:
https://www.githubstatus.com/uptime?page=31
Scaling to human use vs automated/agentic use is a very different thing.
Incredible how reliable the heuristic of "something seems off - probably github being down" has gotten these days
It's big enough that every time it goes down, it surely stops somebody from pushing fix for what they currently have broken, so I wonder if status page services see some kind of ripple from github outages.
About an hour ago I was having trouble browsing repo files in the browser and I thought "A disturbance in the force, is Github down?" Refreshed HN and loaded up their status site. Nada.
(Ofc, in a sensible universe, we just brush that off to a JS/Firefox glitch or my ISP.)
And yet, here I am. My code is not compiling, my AI isn't vibing, nonetheless I can't work! Two more hours before I can get off!
https://www.dayswithoutgithubincident.com
I've been against self hosting internal tools for a long time mainly because of the devops and other overhead. But AI based devops makes it so easy now to spin up whatever you want now that I'm reconsidering that. I use a lot of ansible for several of our deployments. At this point, most of that is managed via codex.
For Git, all you technically need is ssh access and some backup strategy for your server. It would be bare bones but workable. And there are of course plenty of OSS things that are a lot nicer than that.
I'm still using gh and gh actions and we are mostly below the freemium layer with that. But it is kind of slow and honestly a dedicated vm plus some high CPU/memory workers we can spin up on a need to have basis might be a lot faster. With GH outages becoming more common, my hand might be forced a bit.
In recent weeks, I've spun up listmonk (mailing list solution), matrix (as a slack alternative), and a few other things specific to our software stack. A github alternative would be more of the same. We don't need a lot.
The main objection is that with more moving parts to worry about, the workload for me also increases. Things need updating, monitoring, backups, alerting (and responding to alerts), etc. That sucks up my time and that is scarce.
Another reason for self hosting these days is that with agentic AI tools, self hosted things are a lot easier to integrate into agentic systems. If it is self hosted, you don't have to worry about API limitations, rate limitations, walled gardens, etc. All the traditional SAAS silos are becoming a problem from that point of view. The more locked down it is, the bigger the motive for moving away from it. That's why we ditched Slack for Matrix. Slack is hopelessly locked down and tedious to deal with. Matrix is super easy for this.
> For Git, all you technically need is
Technically Dropbox is just rsync.
Also https://xkcd.com/1319/ but for maintenance.
I moved a while back to Forgejo -> https://forgejo.org couldn't be happier. Highly recommended.
Looks good, but I'm not sure about security: https://bearyangry.com/2026/04/29/carrot-disclosure-forgejo-...
Looks lik a terrible source. Like someone ran Claude on the codebase, didn't analyse the results, then vibe coded a blog post. And the dustri.org link doesn't work for me
Anyway. Forgejo's response to it: https://floss.social/@forgejo/116494295922963052
Why do they go down so often? Is it true that the reason is that they've incorporated too much AI without human review?
It's (a) they're under massively increased load because everyone's vibing up new projects these days, (b) they've been in a weird frankenstein "on azure but also we have our own control plane" state for years and they're pushing to no longer have that be the case.
I don't think vibecoding at Github has much to do with it.
Ah, yes. A lot more repos, commits, and most importantly huge PRs.
That makes sense. Thank you!
No, it doesn’t. Their competition is not similarly unstable, despite existing in the same world of LLMs. Think critically.
Devil’s advocate, Pareto heuristic would let us speculate that 80% of LLM traffic would be aimed directly at the largest provider, i.e. GitHub.
Your speculation is that their competitors would naturally not see a commensurate increase in instability while “only” handling 20% of the same crisis?
I don’t buy the excuse. I want to hitch my wagon to those “mysteriously lucky” competitors. (And have. And haven’t had similar issues to Github, since.)
Competitors would be long tail, so a different mode of traffic entirely. Maybe they get spikes that are more easily whack-a-moled than the constant hammering that GitHub receives.
Tough to say as this is all speculative, though.
It's probably a threshold thing isn't it? You wouldn't get 20% of the effect at 20% of the traffic. There's a step function in there somewhere.
I think it’s much more than 80%, it’s probably the default recommendation and folks who aren’t technical would just accept it. Probably closer to 95% or more
Their competition doesn't have nearly the same scale of traffic because they don't have nearly the same scale of users or network effects.
Think critically.
I started using an agent (Codex) on my repo and it went from a a few dozen clones to thousands (3383 this week). I dunno what the agents are doing to clone the repo so many times -- I'm not running 3000 agents or prompts, maybe 10 or so this week. But if this is typical, a 1000x increase in usage across the board can't be good on the system.
GitHub had a blog post about this recently. They reported a significant uptick in volume (repos created, PRs, etc.), which they attribute to AI usage and tooling.
Do you really believe their competition hasn’t seen the same increase? Because their competition certainly hasn’t seen the same instability issues.
This plus in a well-designed system an increase in load might cause new jobs to stop running but shouldn't take down the whole system.
Yes, I truly believe that GitHub is recommended by an LLM orders of magnitude more frequently than any other forge
What competition?
I’ve interviewed a lot of people and when asking about their git experience they’ve said they use GitHub. To a lot of devs they are the same thing.
https://github.blog/news-insights/company-news/an-update-on-...
Microsoft has boasted 30% of their code written by AI.[1] However we could only guess if AI generated code is the issue or something else, or a combination of things.
That being said there was a noticeable trend starting around 2022.[2] That being said they’ve also been doing a big migration to Azure. It’s likely a combination of things.
1: https://www.cnbc.com/2025/04/29/satya-nadella-says-as-much-a...
2: https://www.reddit.com/r/sysadmin/s/LOMPaSv3wY
The instability started well before vibecoding, in around 2018-2019, shortly after the Microsoft acquisition.
https://damrnelson.github.io/github-historical-uptime/
https://news.ycombinator.com/item?id=47591928
This gets posted every time GitHub is down. This chart is not accurate. It is based on data scraped from GitHub's status page and that data is missing historical incidents from the pre-Microsoft era.
Yeah, it’s not even consistent with their own incident history. I spot checked it and consistently found incidents with downtime/elevated error rates in months listed as 100.00000% uptime on that chart.
The unofficial and offical charts are both lying. The GitHub one ignores actual outages and the unofficial ones count minor display bugs in minor features as a “github outage”.
The unofficial one has done that for years though so it’s useful for comparison. If you go back a few years it was regularly at 99.9% uptime.
I personally trigger github actions approximately 50x more than I did prior to AI-driven developer coding and I'm not alone.
Okay so the recent outages are also likely due to increased load due to AI assisted development speeding up workflows.
Totally agree. There's days (or even afternoons) where I trigger more actions than I would have done in a month.
It could be many things. Microsoft mismanaging stuff. Azure. Vibe-coded Github. So much AI slop being committed it adds an extra burden on the servers, etc.
Whilst you're waiting for it to come back, try out AGENT-CI (which is a project I built.), which runs GitHub Actions on your machine: https://agent-ci.dev. (Open source, etc.)
No, it's not like "act," because it uses the standard Github runner, the difference is that the control plane is an emulation of api.github.com, because of this we can do all kinds of nice things:
Caching in ~0 ms. Pause on failure, so you can let your AI agent fix it and retry without pushing.
"Its not like act, because we can add AI"
Is what it boils down to.
> codex "Fix this pipeline, use `act` to verify your changes"
I had extremely bad experience trying to setup act on my Macbook. If this is something that actually works (and doesn't steal my credentials), I'm willing to try it despite AI non-features.
Yea, I've had only barely-success on only a few projects with act. Usually due to steps/scripts that use github-internal APIs, but afaict far from always.
I like that it exists, but what a freaking mess that it's necessary and so difficult to do.
I did not say that, what I said was: It's not like `act` because it's not a rewrite of the runner. It's the standard runner... So the one that actually runs GitHub Actions.
I have tried to use act many times, and many times I've failed.
P.S. pause on failure is also helpful for humans, but I'm trying to be realistic about where the future of programming is going...
What I don’t get about this is how you run OS specific tasks (Windows, macOS, Linux)..
I started playing with proxmox VMs and containers in them (docker and tart) to see if I can build some local infrastructure to properly solve this…
We support macOS via tartlet, but basically it's always linux. If you need windows then it's gonna be an issue.
The jobs runs via containers.
You're affiliated with the project. You should definitely be upfront about that when shilling.
You're right, figured it was implied, but now fixed.
What problem is github solving that has led it to become critical infrastructure for so many? Is it that everyone is remote and VPNs are too much of a hassle to give everyone access to a build server? Is the serving as the authoritative auth for development services? Does it provide better compliance reporting? It just isn't apparent to me what github offers that you can't get elsewhere with at the same cost and effort. I've been in some pretty large orgs with distributed personnel, but this just hasn't ever been a problem.
GitHub solved the original "code collaboration" problem, and now it's a default easy way to outsource repo management. It also has the most integrations. A lot of companies grew up using GitHub.
GitHub was, once upon a time, quite stable. Things have changed: more features, more usage, and automated agents.
I know what it does, but why is it such a problem that Actions is down? I think you did kind of answer it: "A lot of companies grew up using GitHub," i.e., they are using it as infrastructure by default, not because it does something that otherwise can't be done.
It’s well integrated into massively underpriced agentic coding (and noncoding) workflows, I doubt there’s much more reason than that. The hip thing to do now is hold all your docs in github instead of notion so your agent can traverse them locally
I initially thought it was because I ran out of action minute, and was about to upgrade my plan Lucky I came here before hitting the confirm payment button
This is your periodic reminder that Github is growing at ~14x (1400%!) annually. This would be incredible growth for a young, unprofitable, VC-funded startup, even Uber never achieved more than ~3x AFAIK. For a widely-established company that was already very well known and a market leader in its niche for many years? Absolutely unprecedented.
This is a conservative estimate assuming linear growth, the actual number is likely going to be higher. Much higher.
It's not too hard to grow 14X YoY if you start from a hundred customers. If you have hundreds of millions? Yeah, not so easy.
[1] https://x.com/kdaigle/status/2040164759836778878
Are there any GitHub Actions-compatible CI services out there that don't rely on their infrastructure? I know of depot's but no others; are these resilient to these outages or do they still lose functionality? I imagine the latter but I don't know.
github actions themselves can be self hosted, its quite nice actually to be able to keep your same patterns as cloud hosted actions and with one line change to the yaml have it running on your own hardware. I do this for actions that take 6-7 hours so I am not burning through the 3000 minutes that come free with my account.
Self-hosted action runners are not working too right now.
This isn't resilient to this downtime though. Our self-hosted runners are currently not functioning because of some github dependency.
what kind of actions take that long? some kind of compilation task / gigantic test suite ala SQLite?
Founder of Depot here. To my knowledge, we are the first engine to support different syntaxes in this compatible way via Depot CI [0]. Great time to try it out and let us know your thoughts! We’ve built a lot of cool stuff into it like parallel steps, custom images, and a full CLI/API interface so you can literally everything without going into the web app.
[0] https://depot.dev
As someone who partially uses depot but was still affected by this github issue, we obviously haven't moved over enough. We use your runners but github is still blocking us.
Hope you don't mind the public ask, it seems useful for others.
If we're using depot runners, and want to use them directly, or move off of github actions being the controller for when things run: what do you suggest?
Trigger the workflows directly on depot via CLI?
Yes, triggering Depot CI via the CLI is the sure fire way to avoid all dependencies on GitHub.
We’d need more details around what you’re seeing. It is true that if auth across GitHub is broken than we can’t copy your actions out to be used by Depot CI. However, we have a solution in the works for that as well.
In short, Depot CI, our own engine and control plane is not dependent on upstream actions control plane. But still has to listen for commit events to know if/when to run jobs on things like PRs. This to is being removed in the future.
Are you able to bring your own runners? Our org is heavily invested in self-hosted runners at this point and have gotten a pretty tremendous value from it. I think we'd be wise to get away from GitHub's control plane but keep running jobs in our own infra.
Yes, we support this via Depot Managed for all of our products including the latest one: Depot CI [0].
[0] https://depot.dev/products/ci
Is there a tier for open source organizations? Do I have to admin any of AWS that runs behind the scenes or can I pay a fixed price to depot and get it to solve everything out of my way?
I used to use Cirrus CI as an alternative to GitHub Actions and am looking for a new alternative. I wonder if Depot could fit in the same way for my needs. I need to run builds and tests in Windows, Linux and macOS.
there are a couple and have very good reputation - though I've never used them
https://www.blacksmith.sh/ and https://runs-on.com/
They also say that they're much cheaper than github
I think both of these provide nodes that are scheduled using GitHub's control plane. They would also not be working right now.
We currently use external runners (Blacksmith.sh), but that didn't shield us from this as GitHub actions is still the control plane for triggering and monitoring them.
We're now considering Buildkite (apparently they have a GH actions migration tool) or self hosting something (GitLab CI, maybe even Jenkins), as it looks like that would've kept ticking over since we're still seeing webhooks being triggered today during the downtime.
Try Depot CI as well. Supports a GHA syntax but the entire control plane is ours with our own engine.
If you would like less dependence on GitHub for issues and PRs, please check out GitSocial, it stores everything in git itself, making them portable and offline-first.
If you want an alternative to GitHub Actions, you could self-host Forgejo Actions, but I'm not that happy with the design.
I much prefer Woodpecker CI, which is an open source fork of Drone.io. It supports multiple Git backends like GitHub, Gitea, Forgejo, Gitlab, Bitbucket. It supports running jobs locally, on Docker, and on Kubernetes. And there's autoscalers built in for AWS, Hetzner, Linode, Vultr, and Scaleway. There's a bunch of 3rd party plugins (https://woodpecker-ci.org/plugins) for custom integrations. The UX is also very simple, with OAuth used not only for authentication/authorization but also setting up & accessing repos. The system architecture is great, with separate components that run stateless connected to a database, and a custom plugin is any program that takes environment variables and does stdio. The config file is a good balance of ugly YAML and convenience syntax like shell-style parameter expansion variables.
It probably takes less than 15 minutes to install, set up, and run WoodpeckerCI for a small team, so it's not a big investment to try out or host. With the autoscaling plugins it lets you scale your workload up to whatever size. Honestly you could run it on a laptop since it's written Go.
(to clarify for beginners: the config file docs are found in a section called "workflow syntax" (https://woodpecker-ci.org/docs/usage/workflow-syntax) and variable parameter expansion is buried deep in an environment variables page called "string operations" (https://woodpecker-ci.org/docs/usage/environment#string-oper...). poorly organized docs aside, the system itself works well)
Someone said GitHub is racing to the mythical "zero nines of availability" and I love it
Hmm... 88.8888888%?
Jesus, that's both horrible and seems within reach.
Yep, they just need to improve their reliability by 2%!
https://mrshu.github.io/github-statuses/
This page tells a very different story from GitHub own status page. What is different here?
Github measures/reports the SLA of the individual services.
The external page linked above goes the other extreme and considers it a bad status whenever any individual service is degraded.
In reality the majority of people only use 3 or 4 of the core services the majority of the time but since there's no "core services" SLA/uptime the usability of github for the majority of people is slightly obfuscated.
Part of it is that it considers downtime in any of the services GitHub provides as GitHub being down. So if GitHub had 100 different services, and only one of them was down at any given time (but at least one was always down), then it would show 0% uptime.
They've already been well below that over the last 90 days
The main operating model with git is going to go back to decentralized. Setting up and managing something like https://forgejo.org/ is a way better experience than constant interruptions by a faulty service that can't meet demand.
The open source contribution model as we once knew it is dead; you're not going to accept patches from random agents. The risk is way too high. And you can see that increasingly "AI Slop" makes it difficult to be a maintainer of any semblance of a popular repo.
So what's the value? A durable place to store work? hah.
Discovery? That part of Github has always been shitty.
So that leaves.. Github Actions? The thing that is down every other day and has been the subject of a few ~rug pulls~/attempted price hikes that are almost surely coming back?
How to kill a business 101. The brand damage to business and owner is incalculable.
Yeah I'm getting an error where it says account has been suspended. They really are becoming an embarassment
this has happened to me too. i am guessing then it is not a real reason?
`github-actions[bot]` was disabled for some time, if that's the actor which does the checkout in your setup it could be related. FWIW it's back to working now.
Feels like Github Actions is UP should on the front page (when it happens) at this point. Down is no longer front page worthy
If you don't want to self-host Gitea/Forgejo, I recommend SourceHut for private repos and Codeberg for public ones. Happy to answer any questions you might have for either based on my experience!
This is outrageous. Someone go create a Polymarket.
Please don't. These "prediction markets" are a scourge upon mankind.
Just post here when its up. Its easier...
"Microsoft’s GitHub was positioned to win the AI coding race. Outages got in the way" - https://www.cnbc.com/2026/05/22/microsoft-was-positioned-to-...
free service is down again, let's everyone that use the service for free complain again!!! (sorry for the sarcastic comment but i find it crazy how people feel they are entitled when it's free)
EDIT: sorry i meant this rant at the one complaining for the free service not for the paid customers (which is unacceptable)
There are plenty of paying enterprise users that are also affected.
We pay github quite a bit of money and it's down for us too
It’s down for companies too, if your company org is using GitHub enterprise too.
I was actually shocked when I saw what our org pays for Github, not cheap and defo not free
I found it crazy that you haven't discovered that people pay for github
If you take a bit of a closer look, github.com has a "pricing" page
https://github.com/pricing
GitHub Actions outage sparked direct-action, class-action, mass non-action, and widespread dis-satis-faction.
I've started spending each github outage planning our move to an alternative. I guess I'm not alone. Where are you all moving?
We use TeamCity for CI builds, before that Jenkins. Only accessible from the inside of the network.
Even though it's selfhosted and we don't have a dedicated infrastructure team, I don't remember it ever being down in the last 12 years I have been working here.
It's so weird because github used to be known for rock solid stability and now the entire reputation has changed.
You must be new. github was never that stable.
Will more copilot usage fix this? We should try more copilot.
no maybe we should make copilot the pilot so the bad humans in the loop finally cannot break anything.
What could be the cause of GitHub issues from an engineering perspective?
This is great because I finally set up Actions yesterday for a new project of mine and of course it’s failing today and thinking I screwed up the yaml.
I don't understand anyone still using github for anything unless they have to or have payed for it. Move literally anywhere else
I switched to GitLab a while ago and then spun it up locally.
Something’s wrong when my own infrastructure is more reliable than Microsoft’s.
Let us know when your infrastructure sees the load that Microsoft's does and how you've handled it.
Shout out to all my SF 5am crew checking if their overnight prs passed CI. Real 597 “member of technical staff” energy. I guess we should expect this, it is a Tuesday!
My first time using GH Actions was last week. GH was so flaky that pulling a submodule failed >50% of the time. I had to write a script to retry pulling the submodule in a loop.
I've done some hacky shit in CI scripts, but none made me more mad than that one.
I wonder if these github failures are just systematic incompetence or MS cutting budget on purpose to promote its own cicd tools
Or possibly an elevated number of AI Slop Cannons aiming their LLM generated hallucinations at github hosted repos?
Hey at least Copilot AI Model Providers have 100% uptime, so there's that
I have fun somebody imaging somebody internally explaining that this is a heavy traffic page and we should use it to increase reach.
Too many DEI hires? Or maybe H-1Bs? Or maybe it's a vibe coding problem.
I started an open source Git platform. Can be self hosted. I would call it beta at this point if you are interested in trying it. https://velogit.com
I guess a link to the source code would be helpful https://velogit.com/velogit/velogit
And it is bypassing mandatory GHA Pipeline check and giving green. So be careful when merging/reviewing your PRs cause.
List of things "DoS"d by AI:
- GitHub
- Hiring budgets
- RAM (/personal computing in general)
- Electricity
- Media/Content
- Truth
LoL they added "Copilot AI Model Providers" in githubstatus and it has 100% up time.
Thanks for pointing out that nobody is using that thing
here all is ok, 3 actions without problem
I think we should start betting if GitHub will be down on Polymarkets or something at this point.
The future of SRE will be the company putting some amount of money on a prediction market against the site going down and you get to take home the winnings as long as the site stays up.
No way - everyone tells me the AI adoption is going great?
'Degraded' should be banned in status pages. It sounds just irresponsible, like "Yeah, it can be slow or something sometime. Whatever. Who cares"
Straight-up, "degraded" should strictly mean "may be slower, or so slow it randomly fails" on these kinds of status pages.
How would you call "available, but only sometimes"?
The whales are all dying, and we don't know why. Well, some are still alive for now though so maybe it's not so bad...
Does anyone use any good alternatives to GitHub Actions?
oh man spent so much time trying to debug what's going on. I have a complex setup with GitHub Actions and self hosted runners so I thought it's something broken in my CI setup
Ugh, same. 30 mins with 2 devs trying to figure it out before they posted an update.
Now PRs are piling up! https://github.com/mohsen1/tsz/pulls
How's the AI generated code running for ya?
Contingency action plan: Codeberg. Engage.
Has anyone actually moved off? If so where?
I like being able to vote with my (teams) wallet and I'm tired of staying out of convenience
I moved to Codeberg and self hosted Forgejo. I'm happy.
Stop relying on Github.
Self hosted Gitlab with self hosted (or AWS) runners running your pipelines.. We only use Github as a mirror for our public repositories.
Tell Claude to fix it, simple.
i still can't see many pull requests in a bunch of repositories... it's been over a month
Super odd make productivity useless
When is it up?
Github is more likely to be up before noon in UTC timezone. i.e. before the majority of US users are online and causing load.
Or maybe it's before the GitHub internal devs are online and deploying changes.
This has become so typical that we've started working on a modern Github alternative called Plain.
Perfect timing that we post https://www.jxd.dev/writing/building-plain just as this latest incident started.
It should be up again
With the increasing challenges from bots and ai agents created with toddler level clarity, Self hosting is going to continue to work.
my work is totally stop. cry
Zero Nines. Bogus.
microsoft github should work at restoring interop with noscript/basic HTML browsers...
I agree, but that's not at all related to this outage.
Yeah, just reminding people here about that.
I am trying to refrain my "off topic" rants... but such microsoft github abuse is generating so much hate due to their dominant market position, it is hard.
Too many times we've been bitten by this - it has been an issue too many times to count.
This is why we don't use Github Actions, kids.
Seriously, its a proprietary build service that puts the keys to the kingdom in someone elses' control. Just: No!
Print this status page to PDF so you've got it handy next time someone castigates you for not using Github Actions, folks.
So, what do you use?
Another outage at GitHub with actions and pages not working thanks to the AI agents Copilot and Tay.ai creating more issues. Last time this happened was 6 days ago. [0]
This time today it was caused by friendly fire by the automatic suspension of the GitHub Actions bot which is now a "Ghost" user. Since there is no CEO of GitHub to contact it we are just going to see more [1] of this again.
You might need to push a critical change soon, but now you cannot. You won't get any of these issues if you self hosted as I said 6 years ago...[2]
[0] https://www.githubstatus.com/incidents/g6ffrm0rfvz9
[1] https://news.ycombinator.com/item?id=48085501
[2] https://news.ycombinator.com/item?id=22867803
lol
https://github.blog/changelog/2026-05-15-github-app-installa...
I'm guessing related to this? The blog post is dated 11 days ago but I just noticed a blue banner on my actions page today.