We are experiencing this as well. Don't know if Sonoma is related as we've been on it for a while, but having tests time out on us (if they go over 2 hours depending on the "gap"). It's frustrating too because those that time out provide no logs at all for the testing-without-building portion of the workflow. Here is an example of one that we got to pass today, it is a workflow with 3 test components that should run in parallel (UI, Snapshot and Unit), normally it's much quicker but has been timing out / taking a long time for the past couple of days. -- To get these times I downloaded artifacts for each of the test flows, and took the start and end times from xcodebuild-build-for-testing.log and xcodebuild-test-without-building.log You can see that there is a 45 minute or more gap in between when the builds finish and when the tests begin. If you look at the total time for each individual test, it pretty closely aligns with Test Finish - Build start -- Our UI Tests "Said" they took 1h 35m, But if tests were to have started as soon as build finished, it would have only been 36m total. My suspicion is that there is a shortage of test runners (which are often on a different machine than the builders), and so the testing is queued. If this is the case it shouldn't count against our 2h time limit (so they don't time out), and it shouldn't count against our usage.

Hello from 2024, it's happening again for us, started 2 days ago on some builds, now happening pretty reliably (or the opposite) exceeded the maximum allotted time Also: Running ci_post_xcodebuild.sh script failed (exited with code -1) Which is probably a red herring, because we only have a single echo and if statement in that script.

Checking in, did you find an answer? We're approaching this limit, so may end up having an answer (though would like to know what it is before we hit it)

Starting yesterday (About the 1pm or so) and continuing today, we have been having intermittent, but very consistent build failures in Xcode Cloud, failing during the "Fetch Source Code" step, and getting a 502 (like OP) on a random git LFS file retrieval. Our SCM is GitHub, so unlikely this is an issue on the server side. This is causing a random selection of our builds to fail (90%+), with only a few getting through successfully on retry. The error shows as: The action could not complete due to an error. The error may not occur on rebuild. The logs show: From http://github.com/grindrllc/grindr-3.0-ios branch <sha> -> FETCH_HEAD Downloading <Various Files> (1.2 KB) ... (for a random number of files before hitting error) Error downloading object: <Some random file> (<sha>): Smudge error: Error downloading <Some random file> (<sha>): batch response: Fatal error: Server error &{%!!(string=http) %!!(string=) %!!(*url.Userinfo=) %!!(string=github.com) %!!(string=<our repo>/info/lfs/objects/batch) %!!(string=) %!!(bool=false) %!!(string=) %!!(string=) %!!(string=)}s(MISSING) from HTTP 502 Errors logged to '/Volumes/workspace/repository/.git/lfs/logs/.log'. Use `git lfs logs last` to view the log. error: external filter 'git-lfs filter-process' failed fatal: <Some random file>: smudge filter lfs failed Kind of feels like this could be related to 502's, and if they are intermittent, the LFS pull of a git fetch would have a lot. Possible it's some kind of rate limiting?