Hi
When I log the request.completed.count through the :8093/admin/vitals I get a number of 125. But in system:completed_requests I get 14. Are not these numbers supposed to be the same? Am I missing anything?
Thanjs
Hi
When I log the request.completed.count through the :8093/admin/vitals I get a number of 125. But in system:completed_requests I get 14. Are not these numbers supposed to be the same? Am I missing anything?
Thanjs
system:completed_requests will have entries that took more than 1s
vitals will have all the requests cc @marcog
Thanks a lot for your response. But where from does the vital get this information. My understanding was that completed requests only get logged in the completed_requests catalog table or in the audit log if audit log is turned on. I have not turned on audit.
Also, I am not able to understand the below metrics
request.per.sec.15min
request.per.sec.5min
** request.per.sec.1min**
How do I interpret them? In one run, I got all zero for request.per.sec.1min but got values for the other metrics
They are not the same information.
The vital endpoint reports interesting node counters, and the completed.count is the total number of requests processed by that node since the start.
System:completed_requests contains a subset of requests that have completed that are interesting for some reason.
For enterprise editions the system keyspace collates requests across all nodes.
By default we collect the last 4000 requests per node that have taken more than one second, but this can be changed to different limits and different criteria (eg all the requests that have terminated with a specific error, or executed by a specific user, or…)
In terms of request.per.sec, they are exponentially weighted moving averages, meaning that if you haven’t executed requests in the last minute, request.per.sec…1min will be zero.
The monitoring keyspaces are discussed here
The vitals document is discussed here