/area test-and-release
/kind cleanup
/assign adrcunha
Actual Behavior
Tests only run during presubmit, so it's hard to measure how stable they are (e.g., they might fail because of a real bug in the PR). Running the tests on a regular basis (e.g., every 2h) against the master branch will not only give a measure of how stable the tests are, but also if the current code at head is broken.
/area test-and-release
/kind cleanup
/assign adrcunha
Actual Behavior
Tests only run during presubmit, so it's hard to measure how stable they are (e.g., they might fail because of a real bug in the PR). Running the tests on a regular basis (e.g., every 2h) against the master branch will not only give a measure of how stable the tests are, but also if the current code at head is broken.