Graphs missing & get rid of gagues

[sreinhardt]

Interesting, I don't believe I'm seeing this, but I might be running a non-release version. Could you tail the apache error logs please.

tail -n 500 /var/log/http/error_log | grep -i ".png" >> /tmp/httpd-log

Or if you force ssl:

tail -n 500 /var/log/http/ssl_error_log | grep -i ".png" >> /tmp/httpd-log

Please attach these and we'll take a look.

[jwelch]

Did anyone get a chance to look at the initial issue? 'Fixed perfdata graphs to show the proper amount of graphs per page'

[Box293]

I've just checked this on a R1.5 system that I upgraded from R1.4 and it all works fine for me. I've tested with both Highcharts and PNP graphs.

Can you post some screenshots showing the problem please.

[jwelch]

You see 10 graphs on a single page and a way to move to the next page?
On the service I checked, I see 31 graphs on a single page (there are more than 32 volumes returning data)

[Box293]

Yes only 10 a page.

Is this perhaps one service that returns 32 volumes? As apposed to 32 services returning one graph each?

[jwelch]

>Is this perhaps one service that returns 32 volumes?

Yes, the other way wouldn't make sense for this problem. The check uses the NetApp API to pull info on all volumes and alert on any over a set percentage full. It only displays details for problem volumes to cut down on the data load, but returns performance data for all volumes so the usage trend can be graphed. Supposedly, they've added a 10 graph per page limit to the XI performance data page and a way to navigate to more graphs.
On the service I checked, the page keeps loading graphs and stalls out around 32. (no change from previous behavior)

To be clear, I'm looking at the 'Performance Graphs' tab on the 'Service Status Detail' page. For this service (check_netapp_volumes) on this
host, I see 19 graphs and 13 of those annoying little gauges. After the update, I would have expected to see 10 graphs and some method
of navigating to more. The resulting page is too large to capture entirely and I'm not sure if the changes should show up at the top or the bottom.

I just checked and the check output in this case is 1374 bytes so I don't think it's being truncated.

[sreinhardt]

Right, 1.2k should definitely not get truncated last I checked. Also if you were to check with graph explorer you should see the option for all rrd sources per service, if the data is coming into nagios properly. Troy and I are going to set about making a few test cases here and see what we can find out when adding far more than 10 perfdata tracks.

[jwelch]

Actually, I've never used Graph Explorer before. Looking at one of the hosts with lots of graphs, Graph Explorer seems to have the same problem.
The bottom of the page has a navigator control and says 'page 1 of 2', '5 per page', but it's got 25 graphs on the page...some of which have broken
icons.

[sreinhardt]

Can you show us some screenshots of where you are seeing this? At this point in time, my testing shows that using PnP graphs and looking at on both host and service detail pages on the performance graph tabs, the number of graphs are correct for the number of tracks that I have or number of services that I have in the case of host detail. The key difference is that the service detail page does not have the same more graphs option that host detail pages do. We are putting in an internal feature request to extend the same functionality to service detail as well.

As a side note, the changelog note referenced originally was not regarding either of the performance data tabs for host\service detail pages, it was a separate page entirely.

[brandon.pal]

I'm still not seeing page #/# on any of my graph pages. It's just stopping and not showing anything else.