Description

When the number of contacts in the peer table grows large, we may experience performance degradation in key discovery operations. We need to detect and address any bottlenecks.

Affected operations

• Lookup - Finding nodes closest to a target
• Revalidate - Checking liveness of existing contacts
• Prune - Removing stale/unresponsive contacts
• Filtering contacts - Selecting contacts by protocol (discv4/discv5)

Tasks

• Add metrics to measure time elapsed on the affected operations
• Investigate and fix any performance degradation detected

Related issues

• Reintroduce a proper implementation for the Kademlia table #4245 - Reintroduce a proper implementation for the Kademlia table (may help resolve this)
• Cap number of contacts in Kademlia table #3716 - Cap number of contacts in Kademlia table