relay: Improve connect failure cache lifetime with a backoff algorithm
With ticket #24767 (moved), we now have a cache that tracks the relay to relay connection failures. We use that to avoid reconnecting to a relay that we know previously failed.
We could improve here with some sort of backoff algorithm instead of the flat hardcoded 60 seconds waiting period.
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information