Ad Widget

Collapse

Failed to accept an incoming connection: from unknown IP: reading first byte from...

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • marcossilva55
    Junior Member
    • Jun 2021
    • 6

    #1

    Failed to accept an incoming connection: from unknown IP: reading first byte from...

    Hi guys,


    I have a good number of hosts being monitored on my Zabbix where, every 10 minutes, these hosts send a ping to Zabbix to indicate that they are working (UP). What happens is that currently there are many false incident alarms warning that some hosts are down and they are actually up (we can confirm this). (Hosts that are giving false alarms vary, sometimes they warn that they are DOWN, sometimes they warn that they are UP).

    When checking the Zabbix_server logs it is possible to see several logs repeated as in the following print:

    Click image for larger version

Name:	Captura de tela de 2021-06-02 23-38-02.png
Views:	428
Size:	195.2 KB
ID:	426008

    I don't know how to proceed, I've tried several fixes, like setting the Timeout of zabbix_server and zabbix_agent to 5, 10 and 30 and nothing worked.

    I tried to increase the number of StartPollers, StartPollersUnreachable, StartTrappers and StartPingers and apparently the number of active hosts has increased significantly, but it is still far from what it should be.

    If I increase these settings further, the database can't handle the number of connections, resulting in "Too many connections" error.

    Any hints?

    Note: Zabbix is ​​running on the docker on a linux machine.
Working...