exponential backoff on client retry

If there is a problem to connect to the broker or the proxy currently the snowflake client does a constant retry. That can overload the broker (#40465 (closed)). Should we have some increase on the waiting time between retries when there were multiple failures?