Suddenly getting 503 for all HTTP providors - Restarting Traefik restores services

Traefik Traefik v2 (latest)
1

I've had Traefik running since 2020-12-24 with the following static configuration:

[entryPoints]
  [entryPoints.web]
    address = "10.7.7.77:8081"

[entryPoints.web.proxyProtocol]
  trustedIPs = ["10.7.7.107"]

[providers]
  [providers.file]
    directory = "/usr/local/etc/traefik.d"

[log]
  filePath = "/var/log/traefik/traefik.log"

I have a few of *toml files in the /usr/local/etc/traefik.d directory that follow this pattern:

[http]
  [http.routers]
     [http.routers.to-mydomain]
      rule = "Host(`mydomain.tld`, `www.mydomain.tld`)"
      service = "mydomain"

  [http.services]
    [http.services.mydomain.loadBalancer]
      [[http.services.mydomain.loadBalancer.servers]]
        url = "http://mydomain.tld"

It's been working correctly ever since, until now. There is no information in logs, no indication of any OOM kills. The backend HTTP services themselves are also fine. Traefik is still running.

I had to restart Traefik to get things working again.

2

There is an open issue for a similar (if not the same) problem for Traefik 1.x

Am I right in thinking that it isn't recommended to use Traefik in production? Or is there a way to work around this bug?