Add timing data to all Com_ counters in SHOW STATUS
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Percona Server moved to https://jira.percona.com/projects/PS |
Triaged
|
Wishlist
|
Unassigned |
Bug Description
For workload analysis, it would be very helpful to have timing data for all of the Com_ counters in SHOW STATUS.
For example, if you see a spike in query time, which types of commands consumed the time? If a particular type of query is run more frequently, then graphs of the Com_ counters will show that. But if the query is run the same amount of times but has an unstable query plan and isn't using an index sometimes, the counters will be the same and the response time is what will change.
Another example: has the ratio of time spent executing reads -vs- writes changed? Or, how much time is spent executing writes? These answers can be helpful to understand things such as whether the slave is likely to start falling behind soon. Examining a graph of the sum of the response time of all the write queries can show whether there is more execution time than wall-clock time, which will cause the slave to fall behind.
Here's a sample of current output:
Com_show_keys
Com_show_logs
Com_show_
Com_show_
What I'd like to see is
Com_show_keys
Com_show_
And so forth. This should be a floating-point number in seconds. Or, an integer with microseconds elapsed. Either is fine for me.
Changed in percona-server: | |
milestone: | 5.1-11.1 → 5.1-12.0 |
Changed in percona-server: | |
milestone: | 5.1.47-12.0 → 5.1-12.0 |
Changed in percona-server: | |
milestone: | 5.1-12.0 → 5.1-13.0 |
Changed in percona-server: | |
milestone: | 5.1-13.0 → none |
Changed in percona-server: | |
assignee: | Oleg Tsarev (tsarev) → nobody |
status: | Confirmed → Triaged |
Baron,
Please use percona-server to report bugs.