Skip to content

allow configuring gitlabLivenessProbeInitialDelaySeconds

When running on AWS m3.medium instances, the default livenessProbe initialDelaySeconds was actually too short for all the time that chef took on its first run.

That makes sense for how tiny my instances were, but it was pretty tricky to figure out that that's why it was failing! The gitlab logs all looked fine, but the process just died for some reason (kubernetes killed it after 180seconds).

Anyway, I bumped it up to 360 seconds with this patch, and that made it all work nicely. Others may want to do the same.

Merge request reports