Extra logging for errors

[prd-fox]

The DefaultExceptionMapper is providing the stack trace at the debug level.
This means that trying to diagnose issues can be troublesome without knowing where they occurred in the application.

Additionally, the PartyInfoPoller doesn't tell us the URL when a general exception happens. This should be changed to be the same as the ConnectException clause.

We can either:

• increase the level they are logged at
• make it easier to run with the debug logging enabled

[prd-fox]

There can also be too much logging when peers are unable to connect. Although it is useful to report this, it can quickly overwhelm the logs with duplicate info. We should find a way to report the error fully once within a time period, and then a limited amount of info for the same message. See #804 and https://go-quorum.slack.com/archives/C825QTQ1Z/p1561447073127100 for more info.

[chris-j-h]

Another way to reduce the noise when logging errors is to make the distinction between critical and non-critical errors. Being unable to connect to a peer is an error but does not necessarily require immediate concern/action.