Streaming Render Service Metrics

Workspaces with a Professional plan or higher can push a variety of service metrics (memory usage, disk capacity, etc.) to an OpenTelemetry-compatible observability provider, such as New Relic, Honeycomb, or Grafana.

Render does not emit metrics for the following:

• Static sites
• Cron jobs
• One-off jobs

General setup

The following steps must be performed by a workspace admin:

1. From your workspace’s home in the Render Dashboard, select Integrations > Observability in the left sidebar:

   

2. Under Metrics Stream, click + Add destination.

   The following dialog appears:

   

3. Select your observability provider from the dropdown. The dialog updates to display fields specific to your provider.

   If your provider isn’t listed, select Custom. Learn how to connect a custom provider.

4. Fill in the provider-specific fields.

   • See instructions for your provider below.

5. Click Add destination.

You’re all set! Your provider will start receiving reported metrics from Render shortly.

Provider-specific config

When creating a metrics stream for your Render workspace, you provide different information depending on your observability provider:

See details for each supported provider below, along with instructions for other providers. Please also consult your provider’s documentation for additional information.

If there’s a provider you’d like us to add to this list, please submit a feature request.

New Relic

For Region, select US or EU according to where your New Relic data is hosted.

For License key, create a new key with the following steps:

1. From your New Relic API keys page, click Create a key.

   The following dialog appears:

   

2. For the Key type, select Ingest - License.

3. Add a descriptive Name (e.g., “Render Metrics Integration”).

4. Click Create Key.

Honeycomb

For Region, select US or EU according to where your Honeycomb data is hosted.

For API key, create a new key with the following steps:

1. In your Honeycomb dashboard, hover over Manage Data on the bottom left and click Send Data:

   

2. Click Manage API keys.

3. Click Create Ingest API Key.

   The following dialog appears:

   

4. Add a descriptive Name (e.g., “Render Metrics Integration”).

5. Make sure Can create services/datasets is enabled.

6. Click Create.

Grafana

Obtain both your Endpoint and API Token with the following steps:

1. From your Grafana Cloud Portal (grafana.com/orgs/[your-org-name]), click Details for the Grafana stack you want to use:

   

2. Find the OpenTelemetry tile and click Configure.

3. Copy the value of Endpoint for sending OTLP signals (this is your Endpoint).

4. Under Password / API Token, click Generate now.

5. Add a token name (e.g., render_metrics_integration).

6. Click Create Token.

7. Copy the generated value starting with glc_ (this is your API Token).

For more details, see the Grafana documentation.

Datadog

To simplify metrics ingestion with Datadog, Render pushes metrics in Datadog’s native format instead of using OpenTelemetry.

Specify your Datadog site, according to where your Datadog data is hosted.

For API key, generate a new organization-level API key from your organization settings page. You cannot use an application key or a user-scoped API key.

Better Stack

Obtain both your Ingesting host and Source token with the following steps:

1. From your Telemetry > Sources page in Better Stack, click Connect source.

   The following page appears:

   

2. Add a descriptive Name (e.g., “Render Metrics Integration”).

3. Select OpenTelmetry as the Platform.

4. Click Connect source.

   Better Stack creates the new source and redirects you to its details page.

5. Copy your source’s Ingesting host URL and Source token.

Other providers (custom)

Consult this section only if your observability provider isn’t listed above.

Render can push service metrics to your OpenTelemetry-compatible endpoint, if that endpoint authenticates requests via an API key provided as a bearer token in an Authorization header.

If your provider’s endpoint supports authentication via bearer token:

1. Consult your provider’s documentation to obtain your OpenTelemetry endpoint and API key.

2. Specify Custom as your provider in the metrics stream creation dialog, then provide your endpoint and API key in the corresponding fields.

If your provider’s endpoint requires a different authentication method:

1. Please submit a feature request to let us know about your provider’s requirements.

2. You can spin up your own OpenTelemetry collector (such as the official vendor-agnostic implementation). Your collector’s endpoint can receive metrics from Render, then transform and forward them to your provider using whatever authentication method it expects.

Reported metrics

Render streams service metrics that pertain to the following categories:

• CPU
• Memory
• HTTP requests
• Data storage

All metrics use OpenTelemetry JSON format. The first component of each metric’s name is render (e.g., render.service.memory.usage).

Some observability providers transform metric names to match their conventions.

For example, Grafana converts the metric render.service.memory.usage to render_service_memory_usage_bytes.

After you set up your metrics stream, inspect incoming data in your provider’s dashboard to verify how it identifies Render metrics.

See names, descriptions, and included properties for each reported metric below.

Universal properties

All reported metrics include the following properties:

Property	Description
service.name	The name of the service (e.g., my-service). Grafana displays this property as job.
service.id	The ID of the service (e.g., srv-abc123).
service.instance.id	For most metrics, this is the ID of the metric’s associated service instance (e.g., srv-abc123-def456). This is not the case for HTTP request metrics. Everything before the final hyphen is the service ID (srv-abc123), and the final component (def456) uniquely identifies the instance. This value enables you to segment metrics by individual instances of a scaled service, and to identify when a service’s instances are cycled as part of a redeploy.

The following properties are also universal but optional:

Property	Description
service.project	The name of the service’s associated project, if it belongs to one (otherwise omitted).
service.environment	The name of the service’s associated environment, if it belongs to one (otherwise omitted).

CPU

These metrics apply to all compute instances and datastores.

Memory

These metrics apply to all compute instances and datastores.

HTTP requests

These metrics apply only to web services.

HTTP request metrics are not reported per instance.

Render aggregates these metrics across all instances of a given web service. For these metrics, the value of service.instance.id matches that of service.id.

Data storage

Each of these metrics applies to one or more of Render Postgres, Render Key Value, and persistent disks.

History of changes to reported metrics

Date	Change
2025-03-11	Added initial set of reported metrics.