Enable rrd-cache and add more poller processess IT should help.



---- Użytkownik Youssef BENGELLOUN - ZAHR napisał ----

Dear Markus,

 

I don’t believe that is it :

 

 

Before we did the upgrade Monday morning, we were running 8 pollers and polling time was in average around 240s so below the 300s mark.

 

After the upgrade, we started running 16 pollers and polling time went down on average between 120s – 180s.

 

Best regards.

 

 

 

 

Youssef BENGELLOUN - ZAHR - Consultant Expert
Prodware France
T : +33 979 999 000 - F : +33 988 814 001 - ybzahr@prodware.fr
Web : prodware.fr

         

 

De : Youssef BENGELLOUN - ZAHR <ybzahr@prodware.fr>
Date : samedi 8 avril 2017 à 11:10
À : Observium Public Support <observium@observium.org>
Objet : Holes / gaps in graphs

 

Dear Observium community,

 

I’ve been running Observium for over six months, and I have been having this ugly holes / gaps in all generated graphs that look like this :

 

 

 

Here are some stats about our observium VM :

 

 

 

We just upgraded the VM with a few vCPUs and RAM last Monday which helped with the reactivity of the solution :

 

 

 

 

Regarding storage, I know it’s running on 10k SAS SAN. I have been trying to work with my hosting colleagues to do some tweaking on the system level, but I feel that in the end flash storage would be more suitable.

 

In your opinion, what could be the issue that we are facing here ?

 

Thank you in advance for your wisdom.

 

Best regards.