Viewing cluster metrics
Enterprise Edition
Tarantool Cluster Manager is a part of the Enterprise Edition.
In Tarantool Cluster Manager, you can view metrics of connected clusters in real time on the Cluster > Cluster metrics page. The list of metrics that Tarantool exposes is provided in the Metrics reference.
Metrics are displayed one by one. To view a metric, select it in the drop-down list at the top of the page. Then, choose a way to visualize it:
- Chart: a time series chart with the metric values displayed as lines.
- Table: a table where the metric values are displayed as numbers in table cells.
Once you select a metric, TCM starts visualizing its current values, updating them once per second. To pause the visualization, click the button on the left from the metrics selector. To stop the visualization, clear the metric selection.
To view metrics of a specific instance, find this instance on the Stateboard, click its name, and go to the Metrics tab of the instance page.
To allow collecting cluster metrics with external systems, such as Prometheus,
TCM provides HTTP endpoints at /api/metrics/<clusterId>
.
Note
Cluster IDs are shown in the cluster selection dialog that opens when you click Cluster at the top of the left navigation pane.
To access such an endpoint, a request must be authorized with an API token
that has a cluster.metrics
permission on the target cluster.
Below is an example of a Prometheus scrape configuration that collects metrics of a Tarantool cluster from TCM:
- job_name: "tarantool"
static_configs:
- targets: ["127.0.0.1:8080"]
metrics_path: "/api/metrics/00000000-0000-0000-0000-000000000000"
bearer_token: QgMPZ22JZ3uw7n0QTbqYGAQDmNDs1JnTkhaC1OlQzWM3utmpV78b23GG97zp8YE3