Repository navigation
feat(caretaker-evals): add Cloud Run job entrypoint for eval runner - #28727
Conversation
|
📊 PR Size: size/M
|
|
You already have 7 pull requests open. Please work on getting existing PRs merged before opening more. |
Summary of ChangesHello, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request enables the execution of the Caretaker Triage Evaluation Suite in a serverless environment by providing the necessary infrastructure for Google Cloud Run Jobs. It includes a dedicated container definition, an execution entrypoint that handles configuration via environment variables, and a robust synchronization mechanism to persist evaluation results to cloud storage. Highlights
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize the Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counterproductive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here. Footnotes
|
![gemini-code-assist[bot]](/web-api/iframe-proxy?url=https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces a Dockerfile and supporting Python scripts (cloud_runner.py and sync_to_gcs.py) under tools/caretaker-agent to run the Gemini CLI Triage Evaluation Suite as a Cloud Run Job and sync results to Google Cloud Storage. Feedback highlights a potential runtime AttributeError in cloud_runner.py if EVAL_CONFIG parses to a non-dictionary JSON value, and a Docker caching issue in the Dockerfile where cloning the repository without a pinned commit can lead to stale, non-reproducible builds.
2 participants
Summary
Adds the Cloud Run Job entrypoint (
cloud_runner.py), GCS artifact sync helper (helpers/sync_to_gcs.py), and container definition (Dockerfile.evals) for executing the Caretaker Triage Evaluation Suite on Google Cloud Run.Details
tools/caretaker-agent/evals/triage/cloud_runner.py: ReadsEVAL_CONFIGJSON environment variables, executesrun_suite()benchmark runs, and triggers automatic GCS result synchronization.tools/caretaker-agent/evals/triage/helpers/sync_to_gcs.py: Uploads evaluation run artifacts (Markdown summaries and JSON execution logs) to Google Cloud Storage (gs://triage-eval-results/runs/) with explicit UTF-8 headers for rendering on web dashboards.tools/caretaker-agent/Dockerfile.evals: Defines the Python 3.13 container environment pre-baking target workspace dependencies for Cloud Run Job execution.Pre-Merge Checklist