Need more metrics provided by default

[SpokeyWheeler]

At the moment, there aren't really sufficient metrics provided by default. Although this can't provide any application statistics, more generic metrics could be made available - pull requests would be very welcome here!

[SpokeyWheeler]

Metric categories:

CPU Memory Disk User Performance Other

[SpokeyWheeler]

CPU:

• [ ] - length of the ready queue : gauge

Memory:

• [ ] - number of virtual segments : gauge
• [ ] - allocated but not used virtual memory : gauge
• [ ] - per session memory used : gauge
• [ ] - number of buffers per page size : gauge

Disk:

• per table:

• [ ] - row count : gauge

• [ ] - data size in KB : gauge

• [ ] - index size in KB : gauge

• [ ] - number of pages : gauge

• [ ] - row size : gauge

• [ ] - next extent size : gauge

• [ ] - number of sequential scans : counter

• [ ] - number of reads : counter

• [ ] - number of writes : counter

• per dbspace:

• [ ] - free space : gauge

• [ ] - used space : gauge

• [ ] - total space : gauge

• [ ] - number of reads : counter

• [ ] - number of writes : counter

User:

• [ ] - current user count : gauge

Performance:

• [ ] - per pagesize pagreads + bufwrites : counter
• [ ] - uptime : counter
• [ ] - last clear time : counter
• [ ] - current time : counter
• [ ] - checkpoint duration : gauge
• [ ] - per pagesize cache hit ratio : gauge
• [ ] - foreground writes : counter
• [ ] - read-ahead usage : gauge

Other:

• [ ] - free log space
• [ ] - btree cleaner activity

[SpokeyWheeler]

Well, it's not going to be that simple, after all. I need to either write a lot of code to manage the SQL for edge cases and complex SQL, or I need to change the way I process the code. Since I'm post-processing the data anyway, and writing an SQL grammar checker is too hard, I'm going to refactor this to not fiddle with the SQL and just post-process the output.

Joy.

[SpokeyWheeler]

I think this is done! 🎉