( common.runtime.properties ):
| Component | Technology Choice | Purpose | | ----------------------- | -------------------------------- | ----------------------------------------------------------------------- | | | catalog (or graphite , http ) | Druid’s internal reporter. catalog is preferred for modern setups. | | Time-series DB | Prometheus | Stores Druid metrics with efficient labels (service, host, tier). | | Visualization | Grafana | Build per-service dashboards and alert rules. | | Alerting | Prometheus Alertmanager + PagerDuty | Alert on high ingestion lag, unassigned segments, or failed tasks. | | Log aggregation | Loki or Elasticsearch (ELK) | Index Druid logs (e.g., task.log , broker.log ) for debugging. | | Tracing (optional) | Jaeger (OpenTracing) | For deep latency analysis on slow queries. |
( common.runtime.properties ):
| Component | Technology Choice | Purpose | | ----------------------- | -------------------------------- | ----------------------------------------------------------------------- | | | catalog (or graphite , http ) | Druid’s internal reporter. catalog is preferred for modern setups. | | Time-series DB | Prometheus | Stores Druid metrics with efficient labels (service, host, tier). | | Visualization | Grafana | Build per-service dashboards and alert rules. | | Alerting | Prometheus Alertmanager + PagerDuty | Alert on high ingestion lag, unassigned segments, or failed tasks. | | Log aggregation | Loki or Elasticsearch (ELK) | Index Druid logs (e.g., task.log , broker.log ) for debugging. | | Tracing (optional) | Jaeger (OpenTracing) | For deep latency analysis on slow queries. |
