My zabbix implementation constantly has this:
QUEUE OF ITEMS TO BE UPDATED
Delay Count
5 seconds 623
10 seconds 3
30 seconds 76
1 minute 945
5 minutes 378
More than 5 minutes 73
Total:0
more or less. But you get the idea, at any given point in time I am about 1000 or so items behind. I have a number of hosts with a number of items that don't check, or fail to check every so often (UserParameters, so my fault). I have tried increasing both the number of server processes, and the number of agent processes, but I cant seem to get the number of checks over one minute delayed to go down (to zero is my goal).
Is there a magick formula for determining the optimum number of servers and agent processes. Something like
(number of server processes/number of hosts) = [number of agents per host per second] / [number of checks per host] = seconds per host
Or some way to determine how many agents and how many servers I have to have with a given number of checks per host in like a 100 host environment.
QUEUE OF ITEMS TO BE UPDATED
Delay Count
5 seconds 623
10 seconds 3
30 seconds 76
1 minute 945
5 minutes 378
More than 5 minutes 73
Total:0
more or less. But you get the idea, at any given point in time I am about 1000 or so items behind. I have a number of hosts with a number of items that don't check, or fail to check every so often (UserParameters, so my fault). I have tried increasing both the number of server processes, and the number of agent processes, but I cant seem to get the number of checks over one minute delayed to go down (to zero is my goal).
Is there a magick formula for determining the optimum number of servers and agent processes. Something like
(number of server processes/number of hosts) = [number of agents per host per second] / [number of checks per host] = seconds per host
Or some way to determine how many agents and how many servers I have to have with a given number of checks per host in like a 100 host environment.
Comment