Till Rohrmann
Till Rohrmann
This sounds very reasonable. Thanks @slinkydeveloper 🙏
The problem hasn't occurred since increasing the timeouts. Closing this issue as done.
It seems to have happened again: https://github.com/restatedev/restate/actions/runs/18652107682/job/53174722538?pr=3917#step:12:364. @slinkydeveloper could we further harden the test case or is this a real problem?
Maybe https://github.com/restatedev/restate/pull/3285 could help with figuring out whether something is wrong here.
Another instance that happened on a release CI run: https://github.com/restatedev/restate/actions/runs/18655391303/job/53231622011#step:7:442.
https://github.com/restatedev/restate/actions/runs/19312605011/job/55237473233 and https://github.com/restatedev/restate/actions/runs/19237930392/job/54994152559 look suspiciously like duplicates of this issue.
Another instance: https://github.com/restatedev/restate/actions/runs/19573572194/job/56054623314?pr=4052
Yet another instance: https://github.com/restatedev/restate/actions/runs/19630219623/job/56211186381?pr=4053
cc @slinkydeveloper
> I wonder whether this is all in all a better solution, we register just one periodic timer when we get leadership that say every hour scans the invocation list...