GitHub Actions Stability

Over the past few weeks we have seen many problems with GitHub Actions. I know you pride yourself in your test suite and multiple non-user-tested deploys a day. May I remind you that quality still needs to be a part of your pipeline. I am tired of troubleshooting our pipelines only to find out GitHub is down.

Not all of our observed problems make it to https://www.githubstatus.com/ so I am not so sure how trustworthy that is.

We are now regretting moving from our stable Jenkins Builds to Action builds. We have work to do, and actions need to be dependable.

4 Likes

Hi @rlove

This post was originally submitted to our Ecosystem board and I’ve now correctly moved it to our Actions board. It is unfortunate that this got lost in the shuffle, as it’s very clear from your post that you are having a less than stellar experience already.

One thing I’m curious about, without actually discussing the Actions problems you’re experiencing, is whether or not you’re having these troubles with a GHE/S account, or with a standard github.com account?

If your workflows are having trouble within an Enterprise account, I would highly recommend (if you haven’t already) reaching out to our Enterprise Support team directly:

If your concerns are for a non-Enterprise account, it would be helpful to know some details about the suggested lack of dependability in using Actions in general.

Hi @nethgato,

I agree with @rlove: The actual stability of GitHub actions is suboptimal.
The last 5 Days GitHub actions had a lot of outages (for me) … as an enterprise customer I opened issues, but this is no compensation for working pipelines.

As an example: Today I wait for a working GitHub actions service already for more then 6 Hours … 6 Hours where I am blocked … and still no working Pipelines …

GitHub Actions is getting somewhat frustrating for me …

1 Like

This has been the 3rd outage in 2 weeks. It’s also by far the longest outage.

One thing that’s particularly frustrating is that https://www.githubstatus.com/ has no details or post-mortem as to what actually went wrong. The most we get is “there’s a problem”, “we’re looking”, “it’s fixed!”

3 Likes