Default healthprobe values now causes max 120 seconds wait before backend is considered healthy

[fredvd]

@frisi You have added the healthprobe options to the recipe. Any reason why you changed the default Varnish settings from https://varnish-cache.org/docs/6.0/reference/vcl.html#probes in the recipe to have a default treshhold of 8?

The result is that when I restart a backend, it now takes 15s interval * 8 threshold = max 120 seconds before a restarted zope backend is considered healthy again. Varnish defaults to threshold 3, 45 seconds max.

[frisi]

not 100% sure @fredvd

personally i use far lower values as my backend is a haproxy that handles timeouts of the zope instances (that take ~30/60 seconds until they can serve the first request)

health-probe-timeout = 2s
health-probe-interval = 5s
health-probe-window = 3
health-probe-threshold = 2

feel free to change/fix the defaults to whatever makes sense and whatever is considered the default setup for plone.recipe.varnish (i guess varnish in front of a zeo cluster)