We need advice on how to fix metric values.
There a screen from our GUI (upper right corner)
As you can see, value for the graphs ITEMs is 185M. That’s OK.
But when we gather those metrics via rest api or CLI and put them in our Zabbix, we get some shaky graphs.
(left upper corner and bash - blue underline. 1-2-3 metrics at that json are quiet different from other values, and thats why we get shaky graphs.)
For some period of time everything was perfect - graphs in Zabbix and in GUI we equal.
But then something goes wrong with metrics. Couchbase cluster works perfect.
This “shaky” trouble can be found in all buckets and all metrics (disk writes per sec, disk reads per sec. Eg. curr_items_filesystem).
We even upgrade out cluster of 45 machines from 4.5.x to 5.1.1. Still the same.
How we can investigate what is wrong with it?