Support logging internal metrics from `seqcli forwarder` to a secondary Seq instance

[nblumhardt]

Forwarder can be tricky to manage: in particular, it's hard to tell when a forwarder instance has gone offline.

We should:

• Add diagnostics.internalLogServerUri and an API key setting that will turn on diagnostic logging to Seq done!
• Periodically log metrics showing ingest rate, forwarding rate, and various errors