Since https://issues.apache.org/jira/browse/CASSANDRA-11352 changed how the stress summary table is shown, new test runs all show undefined in the graph. The solution to this problem, which is also backward compatible, is quite simple:
For new test runs, store stress summary entries all with lowercase names in benchmark.py
For test runs that are already stored in the DB with mixed case names in the summary table, we do a workaround in the UI of creating a dict entry with a lowercase key. The below picture shows that in the UI, the data object will now have dict keys, where the key has mixed case and lowercase. If they dict key was already all lowercase, we don't modify the dict.
@mshuler can you maybe also push your fix for cstar_perf_regression_monitor.py to this branch?
Since https://issues.apache.org/jira/browse/CASSANDRA-11352 changed how the stress summary table is shown, new test runs all show undefined in the graph. The solution to this problem, which is also backward compatible, is quite simple:
@mshuler can you maybe also push your fix for cstar_perf_regression_monitor.py to this branch?