Try to monitor the Four Golden Signals [1]:
 * for the job queue: number of current webpage analyses, and number of
   pending jobs in the queue
 * for analysis reports: number of reports per status (success, or kind
   of error), time the job spent in the queue before being processed,
   and time it took to analyze the webpage.
 * for the web server: number of requests per status, total time it took
   to service those requests, and total size of the response bodies.

Also, exposes the current timestamp and the timestamp when the server
was started, to be able to compute averages per intervals from above
totals.

The metrics are exposed using Go "expvar" package, so memory stats are
also available.

[1] https://landing.google.com/sre/book/chapters/monitoring-distributed-systems.html#xref_monitoring_golden-signals