RollWorkload with New Config

[twcchu]

Hello!

We are running vector aggregator as a helm chart with configuration through existingConfigMap. When we update the ConfigMap, it gets updated in the Vector pods, but the changes are not applied until the statefulset is restarted manually. We have rollWorkload set to the default true. I don't see the checksum annotation on the pods that should roll the workloads (if I am understanding that configuration correctly). Is there something I am missing that will allow the pods to restart automatically?

[twcchu]

There is a thread on the Discord Server from 2022 that says:

the aggregator chart will redeploy the pods when the configmap holding the configuration changes

but that is not happening for my Vector deployment. Is the only other option to change the rollWorkload setting to false and add the --watch-config argument so that the config changes are applied?

[jszwedko]

There's a workaround in vectordotdev/helm-charts#351 that may be applicable here too.

[jszwedko]

Though there it claims that configmap changes should be causing redeployments.