Michael Bridgen
Michael Bridgen
I see these factors: 1. if the controller is shutting down, but a stack is still in progress after the graceful shutdown period has passed, the (Pulumi) backend keeps it...
To be clear, I think it would be OK to have a longer `leaseDurationSeconds`. It means less sensitivity to failures (it takes longer for a failure to be detected and...
This makes sense as a feature of the operator -- thanks for suggesting it @worawatwi.
This seems entirely reasonable to me 😄 Does this mean you are using this operator with a Flux OCI source?
> We know that switching from `SIGKILL` to `SIGTERM` as the `Pdeathsig` has resulted in a complete drop-off in fail-to-unlock issues, which implies that this is due to the thread...
Kubelet will use SIGKILL without SIGTERM and a grace period, in some circumstances: https://kubernetes.io/docs/concepts/scheduling-eviction/node-pressure-eviction/#hard-eviction-thresholds
Absolutely right, that is not a field of ApplicationLoadBalancer. Oops! Thanks for pointing this out, we'll have to rework that section.
I think just being to run more goroutines would be good enough for now (I'm not clear whether this is the suggestion). Making it scale by sharing work amongst several...
> if they set the controller to replica > 1 each controller will fight for each repo and branch It's a good argument for porting to a controller-runtime `Manager`, so...
I think you'll need to revisit the acceptance criteria, if the point was HA rather than horizontal scalability.