DriftQ-Observability

Prometheus + Grafana stack for DriftQ. Ships with:

• Prometheus scrape config for the broker at :8080/metrics
• Grafana preprovisioned with a Prometheus datasource
• A starter DriftQ Overview dashboard
• Basic alert rules (broker down, high lag, high consumer latency)
• Optional OpenTelemetry Collector (OTLP) for future traces/metrics

Works standalone. When driftqd is not running, the dashboard will just show no data.

---

Quick start

git clone https://github.com/driftq-org/DriftQ-Observability.git
cd driftq-observability
docker compose up -d

Open Grafana: http://localhost:3000 (user: admin, pass: admin). Prometheus: http://localhost:9090

Point Prometheus at your broker

By default, Prometheus scrapes host.docker.internal:8080. If your broker runs elsewhere, edit prometheus/prometheus.yml:

- job_name: "driftqd"
  static_configs:
    - targets: ["<your-host>:8080"]

Restart:

docker compose restart prometheus

---

Files

grafana/
  provisioning/{datasources,dashboards}
  dashboards/driftq-overview.json
prometheus/
  prometheus.yml
  rules/driftq_rules.yml
otel-collector/config.yaml
docker-compose.yml

---

Alerts

• DriftQBrokerDown – Prometheus cannot scrape the broker.
• DriftQHighLag – max(driftq_partition_lag) > 10k for 5m.
• DriftQHighConsumerLatency – max(driftq_consumer_latency_ms) > 5s for 5m.

Wire alertmanager later by adding an alerting section in Prometheus config.

---

Customize

• Add/rename metrics in your broker; update PromQL in the dashboard JSON.
• Add tracing by pointing your apps to OTLP (:4317 gRPC / :4318 HTTP).

---

License

Apache 2.0 – see LICENSE.