Skip to content
ChronactaChronactaChronacta Docs

Monitoring

Operational metrics are exposed on a separate HTTP endpoint configured with CHRONACTA_METRICS_ADDRESS (default 127.0.0.1:9090).

Endpoints

Section titled “Endpoints”
Path Purpose
/metrics Prometheus exposition format
/healthz Liveness probe
/readyz Readiness probe (engine disk/WAL; cluster leader also requires replication_lag ≤ configured max, default 0)

Metrics are operational only. Stream storage APIs remain gRPC-only.

RED / USE catalog

Section titled “RED / USE catalog”

RED (request-driven paths)

Section titled “RED (request-driven paths)”
Dimension gRPC append/read Metric
Rate Requests per second rate(chronacta_append_total[5m]), rate(chronacta_read_total[5m])
Errors Failed RPCs rate(chronacta_rpc_errors_total[5m]), rate(chronacta_auth_failures_total[5m])
Duration Latency chronacta_append_latency_seconds, chronacta_read_latency_seconds (histogram)

USE (resource saturation)

Section titled “USE (resource saturation)”
Dimension Signal Metric
Utilization Active consumers chronacta_active_subscriptions, chronacta_active_projections
Saturation Backlog / lag chronacta_replication_lag_positions, chronacta_projection_lag, chronacta_subscription_lag
Errors Engine / projection health chronacta_engine_ready, chronacta_failed_projections, chronacta_stalled_projections

Disk and WAL pressure: chronacta_wal_bytes, chronacta_active_segment_bytes (compare with node filesystem via node_exporter).

SLO catalog

Section titled “SLO catalog”
SLI Source Target (baseline) Alert rule
Write availability chronacta_engine_ready Ready ≥ 99.9% monthly ChronactaNotReady
Append latency p99 chronacta_append_latency_seconds < 500ms steady-state ChronactaAppendLatencyHigh
HA leadership chronacta_cluster_leader Exactly one leader in cluster ChronactaNoClusterLeader
Replication lag chronacta_replication_lag_positions < 10k positions ChronactaReplicationLagHigh
Projection freshness chronacta_stalled_projections 0 stalled ChronactaStalledProjections
Auth abuse chronacta_auth_failures_total < 1/s sustained ChronactaAuthFailuresHigh
On-disk growth chronacta_wal_bytes + segments < 10 GiB or site policy ChronactaStorageBytesHigh

Numeric baselines and workload tiers: performance-methodology.md. HA RPO/RTO: replication.md.

Key metrics

Section titled “Key metrics”
  • chronacta_append_total, chronacta_append_latency_seconds
  • chronacta_read_total, chronacta_read_latency_seconds
  • chronacta_subscribe_total
  • chronacta_backup_total
  • chronacta_wal_sync_latency_seconds, chronacta_segment_sync_latency_seconds
  • chronacta_auth_failures_total, chronacta_rate_limit_rejections_total, chronacta_resource_rejections_total
  • chronacta_rpc_errors_total{method=...}
  • chronacta_global_position, chronacta_wal_bytes, chronacta_active_segment_bytes
  • chronacta_active_subscriptions, chronacta_subscription_lag
  • chronacta_active_projections, chronacta_projection_lag, chronacta_failed_projections, chronacta_stalled_projections
  • chronacta_engine_ready
  • chronacta_cluster_leader, chronacta_cluster_epoch
  • chronacta_replication_lag_positions, chronacta_replication_lag_by_follower{follower_id}
  • chronacta_replication_append_entries_total, chronacta_replication_append_entries_rejected_total

Disable metrics with CHRONACTA_METRICS_ENABLED=false.

Dashboards and alerts

Section titled “Dashboards and alerts”

Rate limiting

Section titled “Rate limiting”

Enable with CHRONACTA_RATE_LIMIT_ENABLED=true and set CHRONACTA_RATE_LIMIT_RPM (default 600). Limits apply per authenticated username, or per X-Forwarded-For / anonymous key when auth is disabled. Rejections increment chronacta_rate_limit_rejections_total and return RESOURCE_EXHAUSTED.

Example scrape config

Section titled “Example scrape config”
scrape_configs:
  - job_name: chronacta
    static_configs:
      - targets: ['127.0.0.1:9090']

Rule file example:

rule_files:
  - /etc/prometheus/chronacta-alerts.yaml