Monitoring

Monitor both node health and API usefulness.

Signal	Why it matters
Latest checkpoint and lag	Detects stale reads.
Peer count and P2P errors	Detects network isolation.
gRPC latency/error rate by service	Protects backend clients and indexers.
GraphQL latency, status, query cost	Protects public read APIs.
Disk growth and I/O latency	Prevents sync stalls and database corruption risk.

tip

Alert on symptoms users feel: stale checkpoint, high API error rate, and failed transaction submission. Process uptime alone is not enough.