Monitoring

Observability options for MindFry deployments.

Health Check

Use SYS.PING for liveness checks:

mfcli ping
# PONG (uptime: 3600s)

Statistics

SYS.STATS returns core metrics:

const stats = await brain.system.stats()

console.log({
  lineage_count: stats.lineage_count,
  bond_count: stats.bond_count,
  mood: stats.mood,
  uptime: stats.uptime,
})

Key Metrics

Metric	Description	Warning Threshold
`lineage_count`	Active memories	> 80% of max_lineages
`bond_count`	Active bonds	> 80% of max_bonds
`mood`	System mood	< 0.2 (depressive)
`uptime`	Seconds since start	N/A

Docker Health Check

HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
  CMD mfcli ping || exit 1

Prometheus (Future)

Prometheus metrics endpoint is planned for a future release.

Logging

MindFry uses tracing for structured logging:

2026-01-17T10:00:00.000Z INFO mindfry::protocol::handler: Connection accepted from 127.0.0.1:54321
2026-01-17T10:00:00.001Z DEBUG mindfry::protocol::handler: LINEAGE.CREATE key="fire" energy=0.9
2026-01-17T10:00:00.002Z INFO mindfry::dynamics::synapse: Cascade triggered from="fire" neighbors=3

Set log level via MINDFRY_LOG_LEVEL:

error — Errors only
warn — Warnings and above
info — Standard operational logs
debug — Detailed debugging
trace — Maximum verbosity