Currently, PDBs (Pod Disruption Budgets) are being created per cluster with a value of 1 for MAX UNAVAILABLE. This setup, especially with a large number of clusters and numerous shards, leads to prolonged reconciliation times.
One potential improvement could involve parallelizing and expediting this process by creating PDBs per shard within a cluster. In this scenario, each shard would have a 1of MAX UNAVAILABLE of replica.
Is it possible to configure the operator in such a manner?
Hi there,
Currently, PDBs (Pod Disruption Budgets) are being created per cluster with a value of 1 for MAX UNAVAILABLE. This setup, especially with a large number of clusters and numerous shards, leads to prolonged reconciliation times.
One potential improvement could involve parallelizing and expediting this process by creating PDBs per shard within a cluster. In this scenario, each shard would have a 1of MAX UNAVAILABLE of replica.
Is it possible to configure the operator in such a manner?