I wrote an SQL Statement which prints out performance data and implemented it into " Check SQL Database". As it runs ~15 minutes I changed the check interval to 24h and increased the timeout.
So far it works fine.
But it does not generate an rrd graph over this check. I can see that it recognises the performance metrics:
I changed the rules for rrd generation:
|[Configuration of RRD databases of services]
|Consolidation functions:|Minimum, Maximum, Average|
|Step (precision):|86400 sec.|
|RRA configuration:|50.0%, 1, 2880
50.0%, 5, 2880
50.0%, 30, 4320
50.0%, 360, 5840|||
(I only changed step(precision))
I waited 4 days since the last changes, so its definitly not an timing issue.
I also converted the rrds after changing the rules with: “cmk -v --convert-rrds --help”
What am I doing wrong?
Sorry, I wanted to put in more pictures, but I am not allowed to as its my first post…
Thanks for help.
The important setting for such a graph cannot be set from inside CMK GUI.
With the settings from the GUI we can only modify the RRA parameter.
But for 1 value per day you need to change the DS configuration of the RRD storage.
The only DS option is “step” but this only sets the minimum interval in seconds a new value is written to the file.
For 1 value per day you need to change the Heartbeat value to something like 90k seconds.
Here is a good explanation how the RRD storage is working. https://oss.oetiker.ch/rrdtool/doc/rrdcreate.en.html#STEP%2C_HEARTBEAT%2C_and_Rows_As_Durations
With “rrdtool tune Number_of_Objects_in_Trash.rrd -h 1:120000” I was able to change it, and now the graph is generated, thank you!
But the command “cmk -v --convert-rrds” reverts this. Is this something that runs regularly? Can I somehow tell it that these are the settings I need? Thanks again for your help.
Thank you so much, we had the same issue and we solve it changing the heartbeat for all the rrds affected. The hearbeat by default is 8460, it has to be changed to a value > than the seconds in a day, for example: 90000.
The next command needs to be executed for each datasource of the affected rrd. You can see the datasources (ds) doing a rrdtool info /path/to/service.rrd. For a .rrd with 3 datasources you should execute it 3 times.
