Do anyone know which tables are getting impacted due to "Performance warehouse cleanup" jobs which runs everyday at 2:00AM CET to aggregates data from measurement_high table to Measurement_medium and then low.
many tables are affected by the cleanup job.
There's basically three things that happen during this step:
1) Aggregation: affected tables are measurement_*, baseline* and percentiles_*
2) Remove timed-out items: affected tables are as above plus some CI_*, dynamic_measure, incidents
3) Remove unreferenced items: affected tables can be dynamic_measure and many others.
Is there a specific reason why you might want to know the exact set of affected tables? Thanks,
Hi Peter and Babar,
After table partition we started facing lot of issue with the read time measurement for dashboard and our DB team is looking for this information. we already have ticket open in support but unfortunately, did not hear from the support yet.
Thank you, please let me know if you have any query.
There is no long running queries/blockings found in database side when the read spike was high. this issue started occuring post table partition with the same pattern everyday. its starts at 3:00AM CET and resolve by 6:00AM CET.
we are suspecting the stats job and performance cleanup job is overlapped and due to this we are seeing this issue.
Auto gathered stats run everyday at 7:00PM CET.
Advisory tool recommended to do stats gathered on this
: SELECT measure_id, minvalue,
maxvalue,sumvalue, countvalue, measurement_high.timestamp FROM
dynamic_measure_temp LEFT JOIN measurement_high ON id=measure_id WHERE
measurement_high.timestamp>=:1 AND measurement_high.timestamp<=:2
ORDER BY measure_id asc, measurement_high.timestamp asc )”
If you have access to support portal then please refer support ticket SUPDT-42957.