Google Cloud Managed Service for Prometheus

The Google Cloud Managed Service for Prometheus destination exports metrics from a Bindplane pipeline to Google Cloud Managed Service for Prometheus, Google Cloud's fully managed, multi-cloud store for Prometheus metrics. This destination sends metrics only.

Supported Telemetry

Metrics	Logs	Traces
✓

Prerequisites

You need a Google Cloud project and credentials that allow the collector to write metrics:

• A Google Cloud project. Its project ID is set with the Project ID parameter, or detected from the credentials if left unset.
• A service account with the Monitoring Metric Writer role (roles/monitoring.metricWriter), which grants the monitoring.timeSeries.create permission required to write metric data. See Control access with IAM. Grant the role to the service account, not to a user account.
• Network access from the collector to the Cloud Monitoring API endpoint monitoring.googleapis.com:443. When the collector runs on a Google Cloud host, this is typically part of default network access.

When running inside Google Cloud with appropriate access scopes or an attached service account, the auto authentication method discovers credentials from the environment, and no service account key file is required. When running outside Google Cloud, create a service account, assign the Monitoring Metric Writer role, and download a JSON key. See Create service accounts and Create and manage service account keys.

Configuration

Project and Authentication

Parameter	Type	Required	Default	Description
Project ID	string	No	(empty)	The Google Cloud Project ID to send metrics to. Detected from credentials if unset.
Authentication Method	Enum: auto, json, file	Yes	auto	The method used for authenticating to Google Cloud.1
Credentials	string (sensitive)	Yes	(empty)	JSON value from a Google Service Account credential file. Shown when Authentication Method is json. Stored as a sensitive value.
Credentials File	string	Yes	(empty)	Path to a Google Service Account credential file on the collector system. The collector's runtime user must have permission to read this file. Shown when Authentication Method is file.

1. auto uses the collector's environment (such as a Compute Engine service account or GOOGLE_APPLICATION_CREDENTIALS). json takes the contents of a service account key. file reads a key file from the collector system.

Advanced

Parameter	Type	Required	Default	Description
Default Location	Enum	No	us-central1	Google Managed Prometheus requires a location resource attribute. This inserts the attribute if it does not already exist.1
Add Unit and Type Suffixes	bool	No	false	When set, adds unit and type suffixes to metric names.
Cumulative Normalization	bool	No	true	Normalizes cumulative metrics without start times.
Enable Target Info	bool	No	true	Enables the target_info metric, which provides information about the monitored target.
Enable Scope Info	bool	No	true	Enables the scope_info metric, which provides information about the instrumentation scope.

1. Valid values cover Google Cloud regions (for example us-central1, europe-west1, asia-east1) and AWS regions (for example aws:us-east-1).

Sending Queue and Persistent Queue (Advanced)

Parameter	Type	Required	Default	Description
Enable Sending Queue	bool	No	true	Buffer telemetry data temporarily before sending to help ensure data is not lost during a temporary network outage.
Number of Consumers	int	No	10	Number of consumers that dequeue batches. Shown when Enable Sending Queue is on.
Queue Size	int	No	5000	Maximum number of batches kept in memory before dropping. Shown when Enable Sending Queue is on.
Enable Persistent Queuing	bool	No	true	Buffer telemetry data to disk before sending to help ensure data is not lost during network outages or collector restarts. Shown when Enable Sending Queue is on.
Persistent Queue Storage	extension	Yes	file storage extension	The storage extension to use for the persistent queue. Defaults to a file storage extension writing to ${OIQ_OTEL_COLLECTOR_HOME}/storage. Shown when Enable Sending Queue and Enable Persistent Queuing are on.

Examples

Service account JSON credentials

This example sends metrics to the project bindplane-gcp using a service account key pasted directly into Bindplane. The default_location is set to us-central1 so metrics receive a location resource attribute when one is not already present.

apiVersion: bindplane.observiq.com/v1
kind: Destination
metadata:
  name: google-cloud-managed-service-for-prometheus
spec:
  type: googlemanagedprometheus
  parameters:
    - name: project
      value: bindplane-gcp
    - name: auth_type
      value: json
    - name: credentials
      value: '{"type":"service_account","project_id":"REDACTED"}'
    - name: default_location
      value: us-central1

Configuration Tips

• Use auto authentication when the collector runs on Google Cloud with a service account that has the Monitoring Metric Writer role. This avoids managing key material in the configuration.
• Set Default Location to the region nearest your workloads. The exporter inserts a location resource attribute only when one is absent, so existing attributes are preserved.
• Leave the sending and persistent queues enabled so metrics survive temporary network outages or collector restarts before being delivered.

Troubleshooting

Permission denied or 403 errors

Symptoms: the collector logs permission or PERMISSION_DENIED errors when writing metrics.

Solutions:

1. Confirm the service account has the Monitoring Metric Writer role (roles/monitoring.metricWriter) on the target project.
2. Verify the role is granted to the service account in use, not to a user account.
3. Confirm the Cloud Monitoring API is enabled on the project.

Metrics rejected for missing location

Symptoms: metrics are rejected or do not appear, with errors referencing a missing location label.

Solutions:

1. Set Default Location to a valid region so the destination inserts the required location resource attribute.
2. If your pipeline already sets location, confirm it matches a valid Google Cloud or AWS region value.

Dropped metrics during outages

Symptoms: gaps in metrics after a network interruption or collector restart.

Solutions:

1. Keep Enable Sending Queue on so metrics buffer in memory during transient failures.
2. Keep Enable Persistent Queuing on with a configured storage extension so buffered metrics survive collector restarts.

Standalone Destination

apiVersion: bindplane.observiq.com/v1
kind: Destination
metadata:
  name: google-cloud-managed-service-for-prometheus
spec:
  type: googlemanagedprometheus
  parameters:
    - name: auth_type
      value: json
    - name: credentials
      value: '{"type":"service_account","project_id":"REDACTED"}'
    - name: project
      value: bindplane-gcp
    - name: default_location
      value: us-central1

Related Resources

• Google Managed Prometheus exporter (googlemanagedprometheusexporter)
• Google Cloud Managed Service for Prometheus documentation