CI that kept running out of memory
A self-hosted runner ran out of RAM running Rust, CodeQL and Trivy concurrently; fixed by serialising the heavy jobs, capping tool memory and adding swap.
Two copies drifted
One idea kept by hand in two places — schema and migration, dev and prod, app and app.
The self-hosted runner has finite RAM and was running rust, CodeQL and Trivy at once. Took two passes: serialize the heavy jobs, cap CodeQL and Trivy memory, one thread for CodeQL, exclude node_modules and build dirs — and then add 8GB of swap, because the caps alone still crashed it. Also added if: always() to those jobs so one failure stops hiding the others' results.