* Improve readme - what even is this?

Propose words for #2836

* Update README.md

---------

Co-authored-by: Nick Craver <nrcraver@gmail.com>

Fixes the case where we loop from `Connecting` -> `BeginConnectAsync` -> `OnDisconnected` and the next heartbeat triggers a reconnect again due to the backoff thresholds resetting. This ultimately causes the backoff to not be respected.

Here, we're doing 2 things:
- Upping the max exponential backoff window to 60 seconds which better handlers clients at scale and mass DDoS cases.
- Only resets the reconnect counter on the first disconnect, not doing so again until we've successfully reconnected and heartbeated again, so backoffs are properly respected.

…2856)

Additional case handling for #2853, respecting backoff in the disconnected -> resurrecting state case. We need to check backoff in that flow as well to prevent a premature `TryConnect`.

Without this, we don't increase the current count and don't escalate in the exponential retry case, so let's both do that and log that something is happening so we can diagnose from the logs.