Zabbix XXL + Zabbix DB in Docker High Disk I/O?

Discussion in 'Software Stuff' started by Eric Faden, May 14, 2017.

  1. Eric Faden

    Eric Faden Member

    Joined:
    Dec 5, 2016
    Messages:
    97
    Likes Received:
    6
    Hey All,

    So I have a new computer I'm setting up to use as a monitoring platform. It is an AMD based machine (GX-415GA + 8GB RAM + 1 TB Seagate FireCuda). Right now I have Ubuntu 17.04 with Docker (official edge) installed on it. I installed Zabbix XXL w/ Zabbix DB from GitHub - monitoringartist/zabbix-xxl: Dockerized Zabbix - server, web, proxy, java gateway, snmpd with additional extensions. I have several devices setup to monitor via SNMP and 2 Supermicro Proxmox servers via Zabbix Agent + IPMI. I am also running NetData on the Monitoring box and netdata on the two ProxMox servers.

    With this setup (even after trying to tune the DB) I am getting a TON of Disk I/O on the mysql process. It is reporting about 100% Disk Utilization with a ton of I/O wait time.

    Anyone have any ideas where to even start looking?
     
    #1
  2. Jeggs101

    Jeggs101 Well-Known Member

    Joined:
    Dec 29, 2010
    Messages:
    1,468
    Likes Received:
    219
    2 point's i'd like to make
    1 - maybe you're logging too much?
    2 - i'd be appreciative if you could show how you are monitoring the servers. maybe a resource how to guide or two? maybe that'll help diagnose why its making that much io. docker shouldn't cause it.
     
    #2
  3. Eric Faden

    Eric Faden Member

    Joined:
    Dec 5, 2016
    Messages:
    97
    Likes Received:
    6
    1) Yeah... that was my thought....

    2) Mostly just patched together things. A lot of it is via SNMP....
     
    #3
  4. wildchild

    wildchild Active Member

    Joined:
    Feb 4, 2014
    Messages:
    394
    Likes Received:
    57
    Well you'd better check your templates, if pull everything at an interval of 30 seconds...

    Also you should increase the number of spawned pollers, that way zabbix will start doing more things in parallel
     
    #4
  5. wildchild

    wildchild Active Member

    Joined:
    Feb 4, 2014
    Messages:
    394
    Likes Received:
    57
    Oh, and not sure what else runs on this machine, but 8 is probably not enough
     
    #5
  6. Eric Faden

    Eric Faden Member

    Joined:
    Dec 5, 2016
    Messages:
    97
    Likes Received:
    6
    8 isn't enough?
    .... I'm just running a home lab...

    The checks were every 30... I'm thinking about moving graphing to influx. Then I'll make these every 5 or something.... Honestly I'm not sure how much logging in need to do from here. I may just use zabbix for uptime alerting.



    Sent from my Pixel XL using Tapatalk
     
    #6
  7. wildchild

    wildchild Active Member

    Joined:
    Feb 4, 2014
    Messages:
    394
    Likes Received:
    57
Similar Threads: Zabbix Zabbix
Forum Title Date
Software Stuff Zabbix monitoring Oct 6, 2015

Share This Page