Monitoring reference
Lookup catalogs for monitoring HoloMUSH: health-check endpoints, the Prometheus metric set, required PostgreSQL extensions, query-performance metrics, and log locations. For the procedures that use these, see Operations.
Health-check endpoints
Section titled “Health-check endpoints”Both core (default port 9100) and gateway (default port 9101) expose:
| Endpoint | Description |
|---|---|
/healthz/liveness | Process is alive |
/healthz/readiness | Ready to accept traffic |
Prometheus metrics
Section titled “Prometheus metrics”HoloMUSH exposes Prometheus metrics at /metrics.
Connections and requests:
| Metric | Type | Labels | Description |
|---|---|---|---|
holomush_connections_total | Counter | type | Total connections (telnet, web) |
holomush_requests_total | Counter | type, status | Total requests by type and outcome |
Commands:
| Metric | Type | Labels | Description |
|---|---|---|---|
holomush_command_executions_total | Counter | command, source, status | Command executions by name, source, and status (success, error, not_found, permission_denied, rate_limited) |
holomush_command_duration_seconds | Histogram | command, source | Command execution latency |
holomush_command_output_failures_total | Counter | command | Failed to deliver command output to session |
holomush_command_rate_limited_total | Counter | command | Commands rejected by rate limiter |
holomush_alias_expansions_total | Counter | alias | Alias expansion count by alias name |
holomush_alias_rollback_failures_total | Counter | Alias rollback failures (requires manual fix) |
Engine and resilience:
| Metric | Type | Labels | Description |
|---|---|---|---|
holomush_engine_failures_total | Counter | operation | Engine operation failures (access checks, capabilities) |
holomush_circuit_breaker_trips_total | Counter | handler | Circuit breaker activations per command handler |
holomush_circuit_breaker_skipped_total | Counter | handler | Sessions skipped due to open circuit breaker |
holomush_ratelimiter_sessions | Gauge | Current number of tracked rate-limit sessions |
Go runtime and process metrics (go_*, process_*) are also exported automatically.
The audit_projection_lag_seconds metric alerts at > 5s JetStream audit lag.
PostgreSQL extensions
Section titled “PostgreSQL extensions”| Extension | Purpose | Required |
|---|---|---|
pg_trgm | Fuzzy text matching for exits | Yes |
pg_stat_statements | Query performance monitoring | Optional |
Query-performance metrics
Section titled “Query-performance metrics”These pg_stat_statements columns drive the query-performance procedures:
| Metric | Description | Alert Threshold |
|---|---|---|
mean_exec_time | Average query execution time | > 100ms |
stddev_exec_time | Variance in execution time | High variance |
rows | Total rows returned | Depends on query |
shared_blks_read | Blocks read from disk | High = slow |
Log locations
Section titled “Log locations”| Component | Location |
|---|---|
| Core logs | stdout (use log aggregation) |
| Gateway logs | stdout (use log aggregation) |
| PostgreSQL | /var/log/postgresql/ (varies) |