A few engineers on our team reported that without specifying the request connect timeout setting, reconnect on timeout has the old behavior without exponential back off. We should make exponential back off the default, since it is the most sensible reconnect policy.
A few engineers on our team reported that without specifying the request connect timeout setting, reconnect on timeout has the old behavior without exponential back off. We should make exponential back off the default, since it is the most sensible reconnect policy.