Large Cluster Configmap Limits #1874
Description
Hi All,
I recently hit an issue where I was trying to provision a large Clickhouse cluster (1000 shards) and hit a limit on the amount of nodes that could exist in the relevant configmap. After 807 nodes are provisioned the configmap is larger than 1MB in size which is a hard limit in K8S/etcd. I am wondering if there is a workaround or ideas on how to partition the config maps appropriately to work around this issue.