TDEngine

taosKeeper exposes Prometheus metrics on port 6043.

See the taosKeeper reference.

The bundled dashboard uses taos_* metric names and therefore matches taosKeeper's legacy /metrics endpoint. The real-data test used TDengine Enterprise 3.4.2.3 and http://taoskeeper:6043/metrics.

Collection configuration

Create conf/input.prometheus/taoskeeper.toml:

interval = 15

[[instances]]
urls = ["http://127.0.0.1:6043/metrics"]
url_label_key = "instance"
url_label_value = "{{.Host}}"
labels = { job = "taoskeeper" }

curl -fsS http://127.0.0.1:6043/metrics \
  | grep -E 'taos_cluster_info|taos_dn_cpu_taosd' \
  | head

New taosKeeper versions recommend /metrics/v2, which primarily emits taosd_* names and is not directly compatible with the current taos_* dashboard. If you migrate to v2, update the dashboard PromQL and scrape every taosKeeper instance; each v2 instance can hold only a shard of the metrics.