Back to Observability guide

    Grafana · Sample

    StreamFlow — Cluster Overview

    This is the exact dashboard the repo provisions in Grafana on the first docker compose up. Nine panels, visual thresholds, refreshed live below with plausible sample data.

    Live · refresh 3s
    Data source: PrometheusRange: last 15mCluster: streamflow-prod · 3 nodes
    Ingestion Throughput
    GREEN
    330.1K evt/s

    rate(streamflow_events_published_total[1m])

    Publish Latency p99
    GREEN
    3.8ms

    streamflow_publish_latency_ms{quantile=0.99}

    Consumer Lag
    GREEN
    1,226events

    streamflow_consumer_lag{group="orders"}

    Raft Replication Lag
    GREEN
    48entries

    streamflow_replication_lag_entries

    Backpressure Throttle
    GREEN
    0.7ms

    streamflow_backpressure_throttle_ms

    Active Shards
    GREEN
    16

    streamflow_shard_count

    Fsync Latency p99
    GREEN
    2.7ms

    streamflow_fsync_latency_ms{quantile=0.99}

    Batches Committed/s
    GREEN
    1,227batch/s

    rate(streamflow_batches_committed_total[1m])

    Pool Overflow Count
    GREEN
    0overflows

    streamflow_pool_overflow_total

    Visual thresholds (as provisioned)

    Publish / Fsync p99 — < 10ms · 10–50ms · > 50ms
    Consumer lag — < 10K · 10K–100K · > 100K
    Backpressure — 0ms · > 0ms · > 1000ms
    Pool overflow — 0 · > 0 / min · sustained

    Thresholds: publish/fsync p99 <10ms green · lag <10K green · backpressure 0ms green

    Provisioned automatically · 9 panels · 16 alert rules (4 critical / 10 warning / 2 SLO)

    Source: GET /metrics → Prometheus → Grafana · observability guide →

    This is a static render of the shipped Grafana dashboard using plausible sample series — not live production data. Bring it up locally with docker compose -f deploy/monitoring/docker-compose-monitoring.yml up -d and you'll see the same panels, wired to your Prometheus.