Question

You maintain a large Java monorepo where the default CodeQL analysis workflow runs on GitHub-hosted runners. Recently the 'Perform CodeQL Analysis' step began failing intermittently with the runner terminating during the database build, and the logs show the process was killed after consuming all available memory. The team must keep using GitHub Actions for scanning. What is the most appropriate first step to resolve the failures while preserving analysis quality?

Accepted Answer

Switch the workflow to a larger GitHub-hosted or self-hosted runner with more memory to accommodate the CodeQL database build

Answer

Reduce the query suite to 'code-scanning' minimal queries so the analysis consumes less memory during the build phase

Answer

Disable the autobuild step and skip database creation, relying on SARIF upload from a prior successful run

Answer

Lower the scan frequency by triggering the workflow only on a monthly schedule to avoid resource contention