Is your feature request related to a problem? Please describe

In CAS, if number of tenants grew very large, number of segments may start breaching the hard limit for virtual file address and file handle count for a node. This may cause OS process to be killed and subsequently causing node drops.

Describe the solution you'd like

In order to avoid this, we will add monitors for file handles and virtual memory address inside DiskThresholdMonitor. This will require us to add a field for virtual memory address and file handles inside Cluster info. These properties will be refreshed periodically from node stats call along with other disk properties. Once these limits starts breaching a particular threshold limit, we put an index create and read_only_allow_delete block similar to what is currently being done in DIskThresholdMonitor for space. This ensures file handles and virtual memory address are always within limit.

Related component

Indexing

Describe alternatives you've considered

No response

Additional context

No response