dashboards dont load properly on first load

[nozlaf]

since upgrading to 2.0.4 I have had one main issue

Panels are blank when first loading a dashboard

blank panels.png

I need to refresh the page to get them to show up



initially I suspected this was due to the upgrade (i did have some other issues with some of my instances having failed)
so I did a fresh install and moved to all new VM's
same issue

I assumed it was becuase I had only brought one instance online, I have three instances now and not a whole lot of data going in but I'm still having the same issue

load is relatively low

instance overview.png

Issue happens no matter which NLS instance I display the dashboard from



I'm using the Nagios supplied vm's for nagios log server, very standard config

[nozlaf]

sorry can someone move this to the customer support section

[cdienger]

What is the heap usage when you run "curl -XGET 'http://localhost:9200/_cat/nodes?v'" ? Consistent high heap can be indicator of needing more memory or the query/filtering is too complex - what filtering and queries are applied to this? Have you tried it from other browsers? Is it happening with all dashboards? Is there anything logged in /var/log/elasticsearch/ logs ?

Try upping the memory for php to 1024M following the steps in https://support.nagios.com/kb/article/n ... e-611.html .

[nozlaf]

heap query shows

Code: Select all

heap.percent ram.percent load node.role master 
          65          60 0.12 d         m      
          41          60 0.09 d         m      
          65          67 0.33 d         *      

There is nothing of consequence in the logs, (slow logs are empty)

normal log just shows the following event sporadically throughout the day

Code: Select all

 update_mapping [eventlog] (dynamic)

if you think its the heap percentage I can increase the memory on these VM's easily or add additional nodes

[tmcdonald]

Are you familiar with the Dev Console / Dev Tools / Pressing F12 in your browser? If so, it would help to open that, go to the Console tab (or whatever it is called in your browser of choice, just the place that shows JS errors) and visit the page under conditions that cause them to not load. Could be client-side and this could show us what's up.

[nozlaf]

the issue has gone away

I increased the memory to each VM but this did not seem to fix it

I had not applied my license and was running in the 60 day grace period, once I applied the license the banner warning at the top went away and the panels all loaded correctly.

I cant say for certain but I suspect the banner was causing the issue
as I don't have the issue anymore I cant tell if there were issues logged to the console in my browser

[cdienger]

Good catch! I've not seen this behavior before so if you come across it again, please run the dev tools as suggested.

[nozlaf]

Sorry its happening again...
see attached console log

[nozlaf]

Code: Select all

host               ip            heap.percent ram.percent load node.role master name
server1name server1ip           23          64 0.05 d         *      6d0e7f5f-25ec-424e-9db6-f552ea810c3c
server2name server2ip            59          58 0.03 d         m      e56a39d0-ac28-4b8a-b6f9-71e9d50970db
server3name server3ip           55          58 0.01 d         m      2b131b4f-7a1b-4761-b6eb-c2761faf9287

[cdienger]

Is the browser configured to use a proxy that may be blocking or slowing down the transfer of some of the requests/responses? This is logged:
jquery-1.12.4.min.js?ver=2.0.4:2 Slow network is detected. See https://www.chromestatus.com/feature/5636954674692096 for more details. Fallback font will be used while loading: http://mynlsserver/nagioslogserver/medi ... bfont.woff
jquery-1.12.4.min.js?ver=2.0.4:2 Slow network is detected. See https://www.chromestatus.com/feature/5636954674692096 for more details. Fallback font will be used while loading: http://mynlsserver/nagioslogserver/medi ... f2?v=4.7.0

I'd be curious to see what the Network portion of the dev tool says in regards to these two requests.