✓ No known CVEs patched in this version
Topics
+2 more
ReleasePort's take
Light signalRelease v3.0.6 introduces `roc logs --full` for full diagnostic archives and adds AppCapacityProvider resources in CloudFormation and Terraform.
Why it matters: Test the new `roc logs --full` command in dev environments to validate archive completeness; evaluate migrating workloads to use the new capacity‑provider options before your next infrastructure update cycle.
Summary
AI summaryAdded roc logs --full for complete diagnostic archives and capacity provider options in CloudFormation/Terraform.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Feature | Medium |
`roc logs --full` exports complete diagnostic archive for a job. `roc logs --full` exports complete diagnostic archive for a job. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Feature | Medium |
Improved `roc logs` troubleshooting with better parsing and separation. Improved `roc logs` troubleshooting with better parsing and separation. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Feature | Medium |
`AppCapacityProvider` added for CloudFormation to run RunsOn on FARGATE or FARGATE_SPOT. `AppCapacityProvider` added for CloudFormation to run RunsOn on FARGATE or FARGATE_SPOT. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Feature | Medium |
`app_capacity_provider` added for Terraform to run RunsOn on fargate or fargate_spot. `app_capacity_provider` added for Terraform to run RunsOn on fargate or fargate_spot. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Bugfix | Medium |
Improved registration and JIT failure logging with timing details. Improved registration and JIT failure logging with timing details. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Refactor | Medium |
Made runner registration more resilient by leasing work and recovering conflicts. Made runner registration more resilient by leasing work and recovering conflicts. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Refactor | Medium |
Updated agent cache upload path to current AWS S3 transfer manager implementation. Updated agent cache upload path to current AWS S3 transfer manager implementation. Source: llm_adapter@2026-05-21 Confidence: low |
— |
Full changelog
Spotlight
- Added
roc logs --full, which exports a complete diagnostic archive for a job including the DynamoDB job record, RunsOn service logs, CloudTrail events, EC2 console output, and agent logs for attempted instances. - Improved
roc logsjob troubleshooting with better job URL parsing, refreshed job facts, clearer streamed log behavior, and separation between job-specific logs and stack-wide logs. - Made runner registration more resilient by leasing registration work, avoiding duplicate JIT registration calls, and recovering stale GitHub runner-name conflicts when retries happen after partial registration failures.
CloudFormation
- Added
AppCapacityProviderso CloudFormation installs can run the RunsOn control plane onFARGATEorFARGATE_SPOT.
Terraform
- Added
app_capacity_providerso Terraform installs can run the RunsOn control plane onfargateorfargate_spot.
Other fixes
- Improved registration and JIT failure logging with timing details to make GitHub API and retry issues easier to diagnose.
- Updated the agent cache upload path to the current AWS S3 transfer manager implementation.
Release resources
- Upgrade guide: https://runs-on.com/guides/upgrade/
- CloudFormation template: https://runs-on.s3.eu-west-1.amazonaws.com/cloudformation/template-v3.0.6.yaml
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About runs-on
Self-hosted GitHub Actions runners made simple. For AWS. 10x cheaper, 40% faster, and unlimited caching. Best alternative to Actions Runner Controller.
Related context
Related tools
Beta — feedback welcome: [email protected]