| otlp-endpoint | Full OTLP HTTP metrics endpoint, for example https://your-collector.example.com/v1/metrics | no | — |
| otlp-traces-endpoint | Optional OTLP HTTP traces endpoint. Defaults to the metrics endpoint with /v1/traces. | no | — |
| otlp-headers | Comma, newline, or semicolon separated HTTP headers in key=value format | no | — |
| github-token | GitHub token used to enrich traces with workflow job and step metadata | no | — |
| summary-only | Sample locally and write only the job summary without requiring any OTLP endpoint | no | false |
| service-name | OpenTelemetry service.name resource attribute | no | github-runner-telemetry |
| metric-prefix | Prefix used for emitted metric names | no | github.runner |
| sample-interval-ms | Sampling interval in milliseconds | no | 5000 |
| export-timeout-ms | Export timeout in milliseconds | no | 10000 |
| include-network | Record aggregated network throughput metrics | no | true |
| include-filesystem | Record filesystem utilization and throughput metrics | no | true |
| include-load | Record load average metrics when the operating system supports them | no | true |
| enable-job-summary | Append a markdown recommendation summary to the workflow job summary | no | true |
| enable-traces | Export workflow and job trace spans to the OTLP traces endpoint | no | true |
| enable-github-api-enrichment | Use the GitHub API to enrich workflow traces with job and step spans | no | true |
| recommendation-cpu-threshold | CPU p95 threshold that should trigger a larger runner recommendation | no | 85 |
| recommendation-memory-threshold | Memory max threshold that should trigger a larger runner recommendation | no | 80 |
| recommendation-disk-threshold | Disk max threshold that should trigger a larger runner recommendation | no | 85 |
| additional-resource-attributes | Extra resource attributes in key=value format separated by commas, semicolons, or newlines | no | — |