We are having an issue where one of our
We are having an issue where one of our workflows starts failing every single instance whenever a CF build runs on a completely separate (unrelated) worker. It keeps failing until we trigger a re-build on the workflow's worker. This is an extremely annoying issue and very unclear why this would even be happening.....
10 Replies
Hey @ac can you please use /link to link your account. And send a instance id where that happened so that I can investigate?
And also wrangler version
Linked. Instance ID: c8783593-47c4-41ba-a7d2-7357e9e8fb47
The builds are happening via Workers Builds, so Wrangler version is probably not relevant here?
What's the version of wrangler on your package.json?
Worker builds will use that version
Ah ok! 4.14.0 on the worker where the workflow lives (call it "Worker A"), but 4.22.0 on one of the "Worker B" workers where the build is breaking Worker A workflow instances
If you can please bump up both of the versions to the latest and report back. If its still happening I will investigate more.
K, I updated everything to 4.28.1, will report back
Hi @avenceslau , that didn't fix it. A standalone build on Worker B still breaks the workflow in Worker A (both on 4.28.1). New instance id: 1878dbe7-b7f0-4c24-9b40-7d5e268cde05
Thanks if I need anything else I will ping you. We are investigating
Do you have any bindings on Worker B that call connect to either your workflow or Worker A?
Yeah, the workflow that is breaking is a binding on Worker B. It also has a binding to another workflow in Worker B that doesn't break on builds though
If possible DM me your wrangler config for worker B
Actually send me both Worker A and Worker B, thanks in advance.
K, sent