Should we be keeping the health check uncommented?

[Gelmo]

https://github.com/jetrails/magento2-varnish/blob/master/Helper/VclGenerator.php#L155

Currently, we are commenting out most lines in the backend definition/array. In particular, we may want to consider using the health check:

backend default {
    .host = "localhost";
    .port = "8080";
#     .first_byte_timeout = 600s;
#     .probe = {
#         .url = "/pub/health_check.php";
#         .timeout = 2s;
#         .interval = 5s;
#         .window = 10;
#         .threshold = 5;
#    }
}

I've not yet looked into the health check itself. If we determine that the health check should be used, we need to ensure that we modify this from /pub/health_check.php to /health_check.php, as we have Nginx configured to use /pub as the document root for PHP files.

[Gelmo]

Some concerns:

1. What if a client is running Magento AND a non-Magento site with a different hostname, and the non-Magento site is also using Varnish? We'd need to ensure the health check file exists on the other site, and that such a file is actually relevant to that site.
2. If the defined backend (within the VCL file) is HAProxy, HAProxy will forward that to a "random" server, and Varnish is not aware of that. Varnish would need to be configured with multiple backends instead, otherwise, we risk Varnish thinking the backend is down when it may only be one server. I suppose it would be possible to do away with HAProxy altogether, define multiple backends in the VCL, and update the VCL on an interval instead of the HAProxy config. We'd need to do some testing on the impact of reloading Varnish