Hi,

I use collectd on various machines to get telemetry to a central server,
and they all use the network plugin.

Telemetry collection starts before the network is up, which makes sense,
but this means that initially, the network plugin fails to resolve the
server address, and does not retry until the configured ResolveInterval
time.

Lowering that number causes a lot of unnecessary DNS traffic, so that is
not a sensible option, neither is making collectd startup dependent on
network availability.

There should be a separate config option to allow a shorter retry time
after a lookup failure. Also, ideally the ResolveInterval should be the TTL
of the received record.

   Simon