Perform a performance test identical to the baseline, but in parallel, run a script that constantly modifies `PipelineRun` labels (e.g., `oc -n jhutar-1-tenant label --overwrite --all pipelinerun jhutar-test-label=$( date +%s.%N ); sleep 1`). This is designed to induce reconciliation failures in the `pipelines-controller`, testing the hypothesis that such failures cause retries and potential double-counting of the `tekton_pipelines_controller_pipelinerun_total` metric. The test will use a single namespace and 100 concurrency. Compare the PLR and TR counts against the baseline to assess the impact of reconciliation failures on the observed discrepancy.