Job randomly fails on Shared runner in Gitlab CI with exit code 137

We have a job in our pipeline that randomly fails with “Job failed: exit code 137”.
Our builds start from the following image: adoptopenjdk/maven-openjdk11:latest and our job simply executes Maven.
We use shared runners. This happens about half of the time. The other half it works without problems.
How can we debug this problem?

1 Like

We are encountering the same problem, but with completely different images.
It happens with all our builds. Not consistently, but sometimes.

This is very frustrating, because these are false positives. Manually rerunning them fixes it, but of course this is not what you want.

We really have no clue on how to debug this. Can anyone help?