Export metrics

This article describes how to use metrics API to export metrics from a RisingWave Cloud project to various monitoring systems like Prometheus, DataDog, and InfluxDB. The metrics include all major components such as etcd, frontend, compute, compactor, and meta.

By default, a newly created service account isn’t given any permissions. To export metrics, you need to grant your service account the ProjectAdmin role. See RBAC for more details.

Step 1: Generate API key

Generate the API key ID and API key secret in the Cloud Portal. See Generate an API key for details.

Step 2: Get Cloud_HOST

Get the corresponding CLOUD_HOST for your region and Cloud provider from the table below:

Region/CloudProvider	CLOUD_HOST
us-east-2/aws	canary-useast2-mgmt.risingwave.cloud
us-east-1/aws	prod-aws-usea1-mgmt.risingwave.cloud
us-west-2/aws	prod-aws-uswe2-mgmt.risingwave.cloud
eu-west-2/aws	prod-aws-euwe2-mgmt.risingwave.cloud
eu-north-1/aws	prod-aws-euno1-mgmt.risingwave.cloud
ap-southeast-1/aws	prod-aws-apse1-mgmt.risingwave.cloud
europe-west3/gcp	prod-gcp-euwe3-mgmt.risingwave.cloud
europe-north1/gcp	prod-gcp-euno1-mgmt.risingwave.cloud

Step 3: Configure monitoring systems

Choose one of the following methods to configure monitoring systems.

The metrics are formatted according to Prometheus standards. If your monitoring collection mode is compatible with the Prometheus format, refer to the Prometheus section below to configure the collection.

Prometheus

To import the metrics to Prometheus, edit the scrape_configs section in the prometheus.yaml file as follows:

scrape_configs:
- job_name: risingwave-remote
  relabel_configs:
  - target_label: namespace
    regex: {NS_ID}
  - target_label: risingwave_name
    regex: risingwave
  - target_label: risingwave_component
    source_labels:
    - component
  static_configs:
  - targets:
    - {CLOUD_HOST}
  metrics_path: /api/v2/tenants/{NS_ID}/metrics
  basic_auth:
    username: {API_KEY}
    password: {API_SECRET}
  scrape_interval: 60s
  scrape_timeout: 60s
  scheme: https

Notes

• Update the Cloud_HOST according to the region and Cloud provider of your RisingWave cluster.
• Update the NS_ID with the specific UUID of your RisingWave project. To find the UUID, hover over the tag next to your project name.

To import the metrics to Prometheus, edit the scrape_configs section in the prometheus.yaml file as follows:

scrape_configs:
- job_name: risingwave-remote
  relabel_configs:
  - target_label: namespace
    regex: {NS_ID}
  - target_label: risingwave_name
    regex: risingwave
  - target_label: risingwave_component
    source_labels:
    - component
  static_configs:
  - targets:
    - {CLOUD_HOST}
  metrics_path: /api/v2/tenants/{NS_ID}/metrics
  basic_auth:
    username: {API_KEY}
    password: {API_SECRET}
  scrape_interval: 60s
  scrape_timeout: 60s
  scheme: https

Notes

• Update the Cloud_HOST according to the region and Cloud provider of your RisingWave cluster.
• Update the NS_ID with the specific UUID of your RisingWave project. To find the UUID, hover over the tag next to your project name.

To import the metrics to DataDog, use DataDog Agent and DataDog integration. If you have not installed the DataDog agent, see installation guide.

Next, edit the openmetrics.d/conf.yaml file at the root of your Agent configuration directory as follows, and then restart your agents.

instances:
- openmetrics_endpoint: http://{CLOUD_HOST}/api/v2/tenants/{NS_ID}/metrics
  namespace: risingwave
  metrics:
  - .*
  auth_type: basic
  username: {API_KEY}
  password: {API_SECRET}

Notes

• Update the Cloud_HOST according to the region and Cloud provider of your RisingWave cluster.
• Update the NS_ID with the specific UUID of your RisingWave project. To find the UUID, hover over the tag next to your project name.
• The limit for the agent is 2000 metrics per instance. To increase the limit, please contact DataDog support.

To import the metrics to InfluxDB, you need to configure Telegraf first. See instructions on how to use Telegraf to scrape Prometheus metrics and the Prometheus input plugin.

In the Telegraf configuration, you need to fill in the Prometheus input section, see Prometheus for details.

Step 4: Set up dashboards

To set up dashboards with Grafana and Prometheus, please follow the Grafana dashboard build instructions to build a set of dashboards with multiple namespace support.

For other systems, or to generate them programmatically, please refer to the script for generating the Grafana user-facing dashboard and build your own.