Configure GCP Metrics

Complete the following steps to configure GCP Metrics as shown in the following diagram.

Select your platform

Select the platform from the drop-down menu.

Install Grafana Alloy

Grafana Alloy includes an embedded exporter to gather and send metrics to Grafana Cloud.

If you have not already installed Alloy where you intend to run Azure Metrics, click Run Grafana Alloy.
At the Alloy configuration screen, enter a token name. The token displays on the screen and is added to the command for running Alloy.
Copy the command and paste it into the terminal.
Click Proceed to install integration.

Set up GCP service account

To export metrics from GCP, the Google Stackdriver Exporter is used. That exporter uses the Google Golang Client Library, which offers a variety of ways to provide credentials.

To learn how to provide credentials, refer to Google Application Default Credentials.
The roles/monitoring.viewer IAM role contains the required permissions. Refer to the Access Control Guide for details.
For legacy Access scopes, you must use the https://www.googleapis.com/auth/monitoring.read scope.

Configure Alloy

Set up Grafana Alloy to scrape resource metrics within your GCP projects.

Navigate to the configuration file for your Alloy instance.

Copy and append the following to your Alloy configuration file to work with the GCP metrics exporter embedded in Grafana Alloy.

prometheus.exporter.gcp "integrations_gcp" {
        project_ids = [
                "<your-project>",
        ]

        metrics_prefixes = [
                "run.googleapis.com/",
                "cloudfunctions.googleapis.com/",
                "compute.googleapis.com/nat",
                "logging.googleapis.com/billing",
                "logging.googleapis.com/exports",
                "serviceruntime.googleapis.com/quota/",
                "storage.googleapis.com/",
                "pubsub.googleapis.com/subscription",
        ]
}

prometheus.scrape "integrations_gcp" {
    targets    = prometheus.exporter.gcp.integrations_gcp.targets
    forward_to = [prometheus.remote_write.metrics_service.receiver]
    job_name   = "integrations/gcp"
}

Find the value for these placeholders in GCP and replace them with the appropriate values in your Alloy configuration file.

Placeholder Value

project_ids The Google Cloud project ID you want to scrape metrics from.

metrics_prefixes Prefixes for the metric type. Each prefix represents a specific GCP resource, potentially followed by the metric types you want to scrape.
Specify the prefixes you want to scrape metrics from. Examples:
All metrics: pubsub.googleapis.com/
All snapshot-specific metrics: ubsub.googleapis.com/snapshot
All snapshot-speciric metrics and a few subscription metrics:
- pubsub.googleapis.com/snapshot
- pubsub.googleapis.com/subscription/num_undelivered_messages
-pubsub.googleapis.com/subscription/oldest_unacked_message_age

Placeholder	Value
`project_ids`	The Google Cloud project ID you want to scrape metrics from.
`metrics_prefixes`	Prefixes for the metric type. Each prefix represents a specific GCP resource, potentially followed by the metric types you want to scrape. Specify the prefixes you want to scrape metrics from. Examples: All metrics: `pubsub.googleapis.com/` All snapshot-specific metrics: `ubsub.googleapis.com/snapshot` All snapshot-speciric metrics and a few subscription metrics: - `pubsub.googleapis.com/snapshot` - `pubsub.googleapis.com/subscription/num_undelivered_messages` -`pubsub.googleapis.com/subscription/oldest_unacked_message_age`

Add any additional filters to further refine the resources you want to collect metrics from.

Placeholder	Value
`request_interval`	The time range used when querying for metrics. Defaults to 5 minutes when not set.
`ingest_delay`	Offset for the Google Monitoring Metrics interval into the past by the ingest delay from the metric metadata. This automatically adjusts the time range used when querying for metrics backwards. The delay is based on the metadata GCP has published for how long the data can take to be ingested. Defaults to `false` when not set.
`request_offset`	Offset for the Google Monitoring Metrics interval into the past. Defaults to 0 seconds when not set.
`drop_delegated_projects`	Drop metrics from attached projects and only fetch `project_id`. Defaults to `false` when not set.
`gcp_client_timeout`	The time in seconds that the collector should wait for a result from the API. Defaults to 15 seconds when not set.