Colleagues, does anyone use zabbix to monitor kubernetes?
I upgraded to zabbix 6.0 a month ago, and happily went to mount kubernetes monitoring (I understand that there is prometheus, but since I generally have all the monitoring in zabbix, I want to shove the cube there as well). I go here: https://www.zabbix.com/integrations/kubernetes
According to the instructions, first I install zabbix helm, then I fasten cluster monitoring. Everything works, except for one, but important function.
Kubernetes: Get state metrics throws an error
I tried to change the timeout, set a different version of kube-state-metrics (not from zabbix), change the API URL (already as a gesture of desperation, the rest works; I got error 403 when changing), even tried to deploy a completely new cluster with brand new zabbix — same error. Of course, I tried and see what’s in the zabbix git — but there are no changes in this part.
Cube deployed from kubespray, kubernetes version 1.21.5
I can throw here the script code that collects this data, but the question is: does anyone even use this monitoring? I want to understand, is it the radius of curvature of my hands that goes off scale, or is it still a common problem?
Name = Get_Sensors
Type = external validation
Key = sensors.sh
Everything. There is nothing else in the template except for the discovery rule.
JSON in the test is read, so it’s correct. The update interval is set.
Keys in the macro format are set, which means they should be automatically picked up. I tried without them, then parse JSONPath preprocessing and create macros, only got even more confused and deleted both preprocessing and macros. There seems to be nothing to parse, the macros are already set.
Now once again I carefully went through all the tabs in the template and in the detection rule, read the docks, did not find and did not understand where and what to do. The list of sensors can change, the names can change, the number of sensors can be arbitrary, so creating and defining data elements manually is wrong, in my opinion.
What to do next to get the data?
IT colleagues, hello!
I turn to you for help.
Situation. Error in nginx logs
Users of the site complain that the site is not working. Error — Service temporarily unavailable.
What they tried to do.
Change nginx timeouts
proxy_connect_timeout 500;
proxy_send_timeout 500;
proxy_read_timeout 500;
This is not a solution. The error is gone, but the site is not available to the user.
Change prefork settings to recommended Bitrix.
Monitoring is configured on the server. You can watch network load, disks, connections, mysql queries.
The Apache is bent. Some apache parameters reach the maximum and give nginx timeout. What Apache settings to check? Apache is bent because of the php settings. What php settings to check?
Site on Bitrix, vds.
Server hardware configuration:
- OS version — CentOS release 6.2
- Bitrix environment version — 3.01
- PHP version — 5.3.14
- Apache version — 2.2.15
nginx configurations
Virtual host configuration
