Dash0 automatically collects metrics for all Cloud Run services in your connected GCP projects, providing visibility into resource utilization, request performance, and infrastructure health.

Cloud Run metrics help you understand how your serverless containers are performing, identify scaling bottlenecks, and optimize resource allocation to reduce costs.

Viewing Cloud Run Metrics

Navigate to Infrastructure > Google Cloud > Cloud Run to see all Cloud Run services across your connected projects. The view displays:

Service name and region for each Cloud Run service
Key performance indicators including request rate, error rate, and latency percentiles
Resource utilization showing CPU and memory usage
Instance counts indicating current and maximum concurrency

Click any service to drill into detailed metrics and time-series charts.

Dash0 collects the following Cloud Run service metrics from Google Cloud Monitoring. Metrics are stored using their GCP Cloud Monitoring metric type names and can be queried using otel_metric_name in PromQL.

Request Metrics

run.googleapis.com/request_count — Total number of requests received
run.googleapis.com/request_latencies — Request latency distribution (histogram)

Resource Utilization

run.googleapis.com/container/cpu/utilizations — CPU utilization distribution (histogram)
run.googleapis.com/container/memory/utilizations — Memory utilization distribution (histogram)

Instance and Concurrency

run.googleapis.com/container/instance_count — Current number of running container instances
run.googleapis.com/container/max_request_concurrencies — Maximum concurrent requests per instance (histogram)

Network and Cost

run.googleapis.com/container/network/sent_bytes_count — Network egress bytes
run.googleapis.com/container/network/received_bytes_count — Network ingress bytes
run.googleapis.com/container/billable_instance_time — Billable instance time (used for cost calculation)

Note

All metrics are collected at the service level and aggregated across all revisions. To view per-revision metrics, use the revision_name label in PromQL queries.

Common Use Cases

Cloud Run service metrics enable performance optimization and cost control. The following examples demonstrate how to use these metrics to improve service reliability and efficiency.

Identifying Scaling Bottlenecks

High values in the run.googleapis.com/container/max_request_concurrencies histogram combined with increasing request latency suggest your service is reaching its concurrency limits. Consider:

Increasing the maximum concurrent requests per instance
Adjusting the maximum number of instances
Optimizing application code to reduce request processing time

Optimizing Resource Allocation

Compare CPU and memory utilization histograms against your configured limits. If utilization consistently stays in the lower percentiles, you may be over-provisioned and can reduce allocated resources to save costs.

Monitoring Cold Starts

Sudden spikes in high percentiles of run.googleapis.com/request_latencies often indicate cold starts when new instances are launched. If cold starts are impacting user experience:

Enable minimum instances to keep containers warm
Optimize container startup time by reducing image size and dependencies
Consider using startup CPU boost for faster initialization

Cost Analysis

Track run.googleapis.com/container/billable_instance_time over time to understand cost trends. Correlate with request counts to calculate cost per request and identify optimization opportunities.

Correlating with Application Telemetry

Cloud Run metrics become more powerful when correlated with application-level telemetry. Dash0 automatically links infrastructure metrics to:

Traces from your Cloud Run services (via OpenTelemetry instrumentation)
Logs emitted to Google Cloud Logging
Custom metrics sent via OpenTelemetry SDKs

When viewing a Cloud Run service in Dash0, you can pivot to related traces, logs, and application metrics to investigate performance issues holistically.

Tip

Ensure your Cloud Run services emit OpenTelemetry traces with the faas.name and cloud.region attributes set. Dash0 uses these to correlate infrastructure metrics with application telemetry. See GCP Resources for attribute conventions.

Querying Cloud Run Metrics

Use PromQL in the Query Builder to create custom charts and dashboards:

promql

1234567891011
# Request rate by service
sum by (service_name) (rate({otel_metric_name="run.googleapis.com/request_count"}[5m]))

# P99 latency across all services
histogram_quantile(0.99, sum by (service_name, le) (rate({otel_metric_name="run.googleapis.com/request_latencies"}[5m])))

# P99 CPU utilization above 80%
histogram_quantile(0.99, sum by (service_name, le) (rate({otel_metric_name="run.googleapis.com/container/cpu/utilizations"}[5m]))) > 0.8

# Cost trend over time
sum(rate({otel_metric_name="run.googleapis.com/container/billable_instance_time"}[1h]))

All Cloud Run metrics include the following labels for filtering and grouping:

service_name — The Cloud Run service name
revision_name — The specific revision identifier
location — The GCP region (e.g., us-central1)
project_id — The GCP project ID

Alerting on Cloud Run Metrics

Create check rules to alert on Cloud Run performance issues:

High latency: Alert when P99 of run.googleapis.com/request_latencies exceeds thresholds
Error rate spikes: Detect increases in request errors using response_code_class label on run.googleapis.com/request_count
Resource exhaustion: Alert when high percentiles of CPU or memory utilization approach limits
Scaling issues: Notify when run.googleapis.com/container/instance_count hits maximum thresholds

See Alerting for details on creating check rules with infrastructure metrics.

Monitor Cloud Run Services