[tempo-distributed] Rolling update tempo-distributed deployment stucks when replicas is equal worker nodes

[khimick]

When deployments replicas is equal worker nodes rolling update can't start update because of anti-affinity rules and using default kubernetes upgrade strategy.
tempo-distributor deployment manifest:

  strategy:
    type: RollingUpdate
    rollingUpdate:
      maxUnavailable: 25%
      maxSurge: 25%

But let say in tempo-querier manifest strategy is hard coded like this:

  strategy:
    type: RollingUpdate
    rollingUpdate:
      maxUnavailable: 1
      maxSurge: 0

I think, there is a two way:

• change manifest like tempo-querier
• change template for ability to choose strategy in values by the users.