Fix CI timeout by using Docker images instead of building from source - #106
Fix CI timeout by using Docker images instead of building from source#106RonTuretzky wants to merge 4 commits into
Conversation
- Reduce timeout from 20 to 10 minutes (appropriate for Docker-based tests) - Replace local Cargo builds with pre-built Docker images from ghcr.io - Add workflow dependency to ensure Docker images are built before test runs - Support both direct push to dev and workflow_run triggers - Update all log collection to use Docker logs instead of file-based logs - Only build the verification script locally (much faster) This fixes the timeout issue where builds were taking >360 seconds and hitting the 20-minute limit. Using Docker images should complete in <10 minutes. 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Remove all unnecessary build steps and manual service starts - Use docker-compose to start ALL services (including router) - Router image is pulled from ghcr.io, no local building needed - Removed verification script build (not needed for this test) - Much cleaner and simpler workflow matching integration test pattern - Test now focuses only on verifying fast aggregation frequency works The test is now much simpler: 1. Set up .env with AGGREGATION_FREQUENCY=0.3 2. Start all services with docker-compose 3. Wait for EigenLayer setup 4. Verify counter increments at fast rate 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
There was a problem hiding this comment.
Pull Request Overview
This PR addresses CI timeout issues in the fast-aggregation-test workflow by replacing local source compilation with pre-built Docker images, reducing timeout and improving workflow reliability.
Key changes:
- Replaced local Rust compilation with pre-built Docker images from
ghcr.io - Reduced timeout from 20 to 10 minutes to match Docker-based execution
- Added workflow dependency to ensure Docker images are available before test execution
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
| workflow_run: | ||
| workflows: ["Docker CI/CD"] | ||
| types: | ||
| - completed | ||
| branches: | ||
| - dev |
There was a problem hiding this comment.
The workflow_run trigger will execute on any completion of the "Docker CI/CD" workflow, including failures. This could cause the test workflow to run against potentially broken Docker images. The condition on line 20 checks for success, but it would be more efficient to filter at the trigger level using types: [completed] with a more specific condition or use types: [success] if available.
The test was failing because the router was using default 30s aggregation instead of the expected 0.3s. The env var needs to be explicitly passed in the docker-compose override.
|
not relevant anymore |
Summary
Changes
ghcr.ioProblem
The CI was timing out during the build phase. Compilation was taking >360 seconds and hitting the 20-minute workflow limit before tests could even run.
Solution
Use the Docker images that are already being built by the Docker CI/CD workflow. This eliminates the need to compile the router and AVS node from source during testing.
Test Plan
🤖 Generated with Claude Code