Separate the backup pod (whether job pod or long running pod) from the validation DB pod. This would allow a target DB to be customized without impacting build/deployments of backup pod. It would also allow the resources for a recovery test DB to be managed separately from the backup pod (specifically when the backup pod is using the long running option)
Details:
Validation DB pod deployment triggered from within backup pod
Default validation image from postgresql ephemeral image (easier cleanup, less resource consumption)
non-ephemeral customization options ? (for those DB's that are too large for ephemeral)
post validation cleanup to spin validation deployment down to 0 (releasing all resources, incl storage)
Idea:
Separate the backup pod (whether job pod or long running pod) from the validation DB pod. This would allow a target DB to be customized without impacting build/deployments of backup pod. It would also allow the resources for a recovery test DB to be managed separately from the backup pod (
specifically when the backup pod is using the long running option)Details: