Add more metrics

[absorbb]

Problem

We have to little metrics to be sure that application is healthy and there are no regressions after upgrading to a new version

Solution

For every significant event that take place during lifetime add prometheus metrics:

• app start / shutdown
• unit initialization (destination, source, api_key, etc) success | error
• new unit created (destination, source, api_key, etc) success | error
• unit test connection (destination, source) success | error
• task_executor.go: source(collection) sync run success | error | cancel | stale?
• uploader.go: api token file processed (break by destination & table) success | error
• backend auth status success | login_failed | error
• open API endpoint statuses http codes? maybe already collected by gin
• mail send status success | error ?
• slack notification send status success | error
• replay status success | error
• ...