These panels had a few copy/paste mistakes, most notable the PDB active
write pool active connections was actually a duplicate of active read
connections.

Add command to extract metrics archives to README

Fix panels related to PDB connections

This commit adds panels for:

* HA sync times
* Processing times
This is a composite panel that replaces the individual ones for
catalogs, facts, and reports
* Processing rates and counts
* GC times and counts

Prior to this commit, this metric was not being emitted from archive
metrics because I had confused jetty's queueSize metric with PDB's
queue_depth.  This commit emits the correct metric for the dashboard to
pick up.

(SUP-3250) Add HA and other PDB panels

Fix queue_depth metric in starlark processor

Add panels for Puppet server GC times and counts

This commits adds panels for PDB JVM garbage collection, not to be
confused with the panels we already have for PDB GC.  I renamed those to
hopefully avoid confusion.

This makes field names like gc-stats_PS_MarkSweep_count consistent
between archive metrics and telegraf.

Add PDB JVM GC panels

This commit adds two panels for route-id mean duration and counts.

Add metrics from route-ids

Prior to this commit, the provisioning datasource file had a 'before'
metaparam on the Grafana service, which caused the service not to
refresh when the file changes.  This commit fixes this by changing the
'before' to a 'notify'.  The 'require' on the Grafana install class
still ensures proper ordering.

(SUP-3319) Refresh service when datasource changes

This commit moves the section on importing archive metrics to its own
file and adds more detail about the process.

(SUP-3329) Add metric archive info to ARCHIVES.md

Release prep v1.1.0