The rrdtool info dump for the server 02 shows that the datatype is gauge but the perfdata.log says that is is a counter.
I am going to guess that the rrd file for server 02 was created a while back and and when it was created, it was created as the datatype of gauge and that is why the graph it correct.
Without having detailed debugging files for when the rrd files are created, it is just a guess as to why that happened.
Without creating a template for those checks, the data format will default to a counter which will not graph correctly.
Be sure to check out our Knowledgebase for helpful articles and solutions!
Ok so I recreated the checks using the wizard. This time I used a label called err_count and events. Now each graph is of type gauge.
However now I just need to figure out how to make a differential graph (Delta graph if you will). So I can see the change in count between monitoring points instead of the accumulated counter.
I haven't tried graphing the data between the checks. You could template it and select a different data type and see if that is what you are looking for.
Take a look at this link the describes what the types are. http://oss.oetiker.ch/rrdtool/doc/rrdcreate.en.html
Be sure to check out our Knowledgebase for helpful articles and solutions!
Ok so I've broken down and created my own custom check script to retrieve the current value, and create a delta from a tempfile it creates. Yay! Have not created a template, but that's next on my list.
However pnp? is still doing something funny to these real values, I'm assuming it's to account for time differentials. So instead of just showing my direct value in integers, they always appear as a float. Any idea how to fix that?