Zabbix server exposed by Traefik v2 causes SSL_read() timed out

[richirojoli]

Hi all,

I am using Traefik v2 as a reverse proxy in front of my docker based Zabbix server, the web works perfectly but the agents cannot connect to the server through a tcp router.

Here is the docker-compose section for the zabbix server:
version: '3'
services:
zabbix-server:
image: zabbix/zabbix-server-pgsql:alpine-5.0-latest
container_name: zabbix-server
restart: unless-stopped
networks:
- zabbix
- traefik
environment:
DB_SERVER_HOST: "pg-zabbix"
POSTGRES_USER: "xxx"
POSTGRES_PASSWORD: "xxx"
POSTGRES_DB: "xxx"
ZBX_ENABLE_SNMP_TRAPS: "false"
labels:
- "traefik.enable=true"
- "traefik.docker.network=traefik"
- "traefik.tcp.routers.zabbixserver.entrypoints=zabb ix"
- "traefik.tcp.routers.zabbixserver.rule=HostSNI(`za bbix.example.com`)"
- "traefik.tcp.services.zabbixserver.loadbalancer.se rver.port=10051"
networks:
traefik:
external: true
zabbix:

In the Traefik configuration I have the zabbix entrypoint like so:
entryPoints:
zabbix:
address: ":10051"

The agent is using zabbix.example.com as the address to send active checks.

Is there anyone who uses Traefik in front of the whole Zabbix configuration? Any Idea why it is failing?