Monitor Neo4j
Select the intended project/database and time range before interpreting DB-Metrics. Use the following columns to review database health and performance.
| Column | Meaning |
|---|---|
| Service Name | Host and port of the Neo4j instance |
| DB Exporter | Whether metrics are being collected (Online/Up or Offline) |
| Service Role | Primary/Writer (the leader), Primary, or Secondary |
| DB Version | Installed Neo4j Enterprise version |
| APOC Version | Installed APOC Core version; must be compatible with the deployed DB Version |
| DB Size | Database size on disk |
| Active Transaction | Transactions running now |
| Cache Total Faults | Page-cache misses that went to disk |
| Bolt Connections | Active client connections |
APOC Core must be compatible with the installed Neo4j release. Check both reported versions after upgrade rather than treating a stale field as conclusive evidence of a package mismatch.
Signals to investigate
- Database size nearing capacity: plan storage before a full filesystem stops service.
- Sustained active transactions: review workload and long-running queries.
- Unexpected leader changes: inspect the former leader and peer connectivity at the event time.
- Elevated cache faults: investigate the working set, memory settings and storage latency together.
- Missing Bolt connections or unavailable exporter: separate client access from metrics collection.
After node removal, check topology and scrape-target updates. Home and host-level observations are covered by project monitoring. For empty/stale panels use monitoring troubleshooting.