Description
Is your feature request related to a problem ?
When sharding prometheus in the high-availability set-up, at least one prometheus pod per shard should always be running to avoid missing scrapes. PodDisruptionBudgets are helpful here
However, the PodDisruptionBudget created by this chart has label selectors that match on every prometheus pod. So if you're running with two shards each with two replicas for example, then you can't guarantee that only one pod per shard is disrupted at a time
Describe the solution you'd like.
When sharding is enabled, create a PodDisruptionBudget for each shard. The existing operator.prometheus.io/shard label could be used for the label selector
Describe alternatives you've considered.
Creating custom PodDisruptionBudgets instead of relying on this helm chart is always an option but its convenient if this chart handles this instead
Additional context.
I can implement this change