During the load testing, we added a wider scenario for the tests, so now we have two tests:

• simple scenario: login -> create wkspc via API -> access workspace URL (will start the wkspc) -> wait for IDE to be ready -> go to Dashboard -> stop wkspc -> delete wkspc
• complicated scenario: login -> create wkspc via UI (select sample) -> wait for IDE to be ready -> test LS -> test commands running -> test debug -> go to Dashboard -> stop wkspc -> delete wkspc

When I was running a complicated scenario, it was failing some time on an issue that could be probably identified in browser logs as ERR_INSUFFICIENT_RESOURCES. It has different consequences on error messages shown to users (see pictures). It does not happen when I was running a simple scenario at all.

The weird thing is that this issue is reproduced only when running load tests, it was not reproduced while running the devfile test suite (running all devfile samples but only by one user and one by one).

I've tried to increase the resources available for the server and also for the pods with tests. It doesn't seem to help much. Maybe would be worth investigating with proper monitoring turned on. Also would be nice to be able to grab HAR from the browser when the test fails.

Note: There is also a lot of WebSocket connection is closed. messages in browser logs, which is probably related to this issue: https://github.com/eclipse/che/issues/19339