Some pods / jobs fail when deployment takes a long time

[Moritz72]

Current Behavior

For me sometimes when deployment happens to take a longer time keycloak crashes or terraform throws an error that the azurite-init job is not in complete state although the logs say it is. This behaviour is very inconsistent and only happens sometimes.

Expected Behavior

The deployment consistently works without issues.

Steps To Reproduce

Even for me this is not easily reproducible and seems to heavily depend on how fast which pods are created.

But after adding

depends_on = [kubernetes_deployment.keycloak]

in the miw deployment and

timeouts {
  create = "5m"
}

in the azurite-init job the deployment worked consistently again.