Reset before observing

This release includes a change that calls .Reset() on each of the process metrics before observing in order to remove older labels. This way we don't have transient or dead processes that stick around in the metrics permanently.

Updates For SQS

Adds metrics for:

• database lock info
• connected clients
• latency probes
• disk usage
• partition counts

Updates For SQS

Updates to fit our use-case in SQS

* update dependencies
* change output to be logs with timestamps
* remove clusterfile creation code to prevent accidental creation