adding a max_fails parameter to upstream member[s]

[ghost]

As defined in nginx upstream documentation: http://nginx.org/en/docs/http/ngx_http_upstream_module.html#upstream

I added this optional parameter because we folks have to deal with a finicky backend that randomly timeout on a single request when under heavy load. When this happens, the upstream gets disabled for 10 seconds (fail_timeout) and for that period of time other upstreams receive the requests which doesn't play nice with our application.

Increasing this parameter to max_fails=3 ensures that load is switched only when really necessary.

  nginx::resource::upstream { $upstream_name:
    ensure               => present,
    members              => $upstream_servers,
    upstream_cfg_prepend => {
      'ip_hash'   => '',
      'keepalive' => '128',
    },
    upstream_max_fails   => 3,
  }

becomes s:

upstream backend {

  ip_hash ;

  keepalive 65;

  server     server1.domain  fail_timeout=10s max_fails=3;
  server     server2.domain  fail_timeout=10s max_fails=3;
}

Let me know if I should change anything, thanks

[jfryman]

Great! Thanks for the code!