Test TestMetricsAndPprofExist
intermittently fails due to Tracee timeout and subsequent "already running" error
#4487
Labels
Description
The
TestMetricsAndPprofExist
test is experiencing intermittent failures related to how the Tracee process is managed. The failure manifests in two stages:Stage 1: Timeout
Initially, the test fails because Tracee does not start within the defined timeout you can view more in issue #4486
Stage 2: "Already Running"
When the test is rerun immediately after a timeout failure, it fails again, but this time due to the error "tracee is already running" (testutils.TraceeAlreadyRunning). This is because the Tracee process from the previous, timed-out run is still active in the background and not properly terminated by the test framework.
The core issue is that when TestMetricsAndPprofExist times out, the Tracee process is left orphaned. The test does not have a mechanism to gracefully terminate a timed-out Tracee instance before rerunning. Currently, manual intervention with
pkill
is required to kill the lingering Tracee process before the test can be successfully rerun.pgrep tracee kill [output of pgrep]
Output of
tracee version
:Output of
uname -a
:Additional details
The text was updated successfully, but these errors were encountered: