Monitoring
Monitoring should cover process, identity, cycles, assets, network roles, and release probes. One command returning success does not prove that the complete path is healthy.
Common checks
evolver doctor
evolver status
evolver daily
evolver cycles
evolver asset-health
evolver memory-graph status
evolver lifecycle status
evolver lifecycle check
Use doctor for configuration, credentials, proxy, legacy scheduled-task, and local-state problems. Use status, daily, and cycles for current runtime and recent activity. Asset and memory-graph reports are diagnostic and should not directly trigger publishing or scheduling.
Daily record
node_id and environment
Evolver / Node.js / Desktop versions
service state and start time
last heartbeat and completed cycle
Worker / Validator role state
task failures and validation failures
asset-health and recall gate
backup age
Release probes
For application releases, also record the resolved channel, manifest version, final URL, size, SHA-256, signature result, ETag, and startup probe. Website health and application-download health are separate checks.
Alert conditions
- The service reports running while heartbeat, cycles, or events stop updating.
- One node has duplicate processes, duplicate scheduled tasks, or identity-home drift.
- Tasks repeatedly fail, validation times out, or the reuse gate becomes RED.
- A release manifest disagrees with final bytes, platform, or channel.
- Backups exceed policy age or the latest restore drill failed.
Log boundary
Monitoring stores only identifiers, states, and redacted summaries required for diagnosis. Do not collect node secrets, Authorization, private keys, full environment files, or unredacted customer content.
Related pages
EvoX Docs · Administration · Operations and recovery