metric-collector-for-apache-cassandra
metric-collector-for-apache-cassandra copied to clipboard
Read Path Dasboards V1
Add the "Read Path Dashboard" as defined in #27.
To be reviewed / tested and merged!
@tjake, Do you know why we would have
mcac_table_key_cache_hit_rateavailable, but notmcac_table_row_cache_hit_rateormcac_table_counter_cache_hit_rate?
I'm guessing these only show up when you enable them in cassandra.yaml
So after looking at MBeans available through jconsole, I can see that there is some inconsistencies for caches at table level:
Also the counter cache is not present at all there (nor chunk cache - but on this one I'm not sure if the table level is really relevant).
@tjake do you think it's relevant to report this inconsistence at the project level?
Other than that, I can use those metrics - at the node level:
Also, at the OS level (collectd), I'm looking for those options (that were available on dashboards we did for Datadog - and are available there out of the box). But I'll look a bit more I did not take the time to do that search seriously yet. Just if you know about it, I'd like to know :).
- Disk Read Latency (io.r_await?)
- Disk Read Queued (io.rrqm/s)
- Disk Utilization (io.util%)