I try to use the following chart for our clusters (on-premise and Azure Cloud) https://git.zabbix.com/projects/ZT/repos/kubernetes-helm/browse.
Kubernetes version is v1.22.6.

The configuration is default as from the original chart except for the zabbix-agent env. The zabbix-proxy is disabled. Container image: zabbix/zabbix-agent2:ubuntu-6.0.5.

The following issues have occurred:

1. When I deployed the chart with the following env:

env:        
  - name: ZBX_SERVER_HOST          
    value: proxy.example.com      
  - name: ZBX_METADATA          
    value: DEV-2        
  - name: ZBX_SERVER_PORT          
    value: 10051        
  - name: ZBX_PASSIVE_ALLOW          
    value: true        
  - name: ZBX_ACTIVE_ALLOW          
    value: true        
  - name: ZBX_DEBUGLEVEL          
    value: 3        
  - name: ZBX_TIMEOUT
    value: 20
  - name: ZBX_REFRESHACTIVECHECKS          
    value: 90

Result: the following errors come every 1- seconds according to the  default livenessProbe in zabbix-agent deployment:

cannot accept incoming connection for peer: <node IP>

However, agents are working fine the push provides the data to the proxy.

1. When I deployed the chart with the following env:

env: 
  - name: ZBX_SERVER_HOST 
    value: proxy.example.com 
  - name: ZBX_METADATA 
    value: DEV-2 
  - name: ZBX_SERVER_PORT 
    value: 10051 
  - name: ZBX_PASSIVE_ALLOW 
    value: false 
  - name: ZBX_ACTIVE_ALLOW 
    value: true 
  - name: ZBX_DEBUGLEVEL 
    value: 3 
  - name: ZBX_TIMEOUT
    value: 20
  - name: ZBX_REFRESHACTIVECHECKS 
    value: 90 

Result: the pods could not start due to startupProbe. This problem also happens with deployment to Azure AKS and no matter which envs are used. 

If in both cases I delete all probes after or before the deployment everything works smoothly. 

Expected:
The deployment should be with probes without any error.