You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When we have a reservation on a relay, and we lose connectivity to that relay, we should make an effort to reconnect to the same relay before searching for other relays to make a reservation on.
This is because making a new reservation will change our multiaddrs which impacts things like DHT provider records which will now be out of date just because our WiFi was a bit flaky for a time or we closed our laptop lid.
The text was updated successfully, but these errors were encountered:
Would a LRU type data structure help get better relays historically?
Furthermore a weighted data-structure (possibly a learning/AI model) which scores the relays based on their reliability/data-structures/latency etc, (prolly overkill) but would be better?
I think this relates to #1953 - i.e. as a general rule, if we've successfully dialled the peer recently, but we are no longer connected to them, we should prioritize dialling those peers and conversely if we have dialled a peer but failed recently, then those should be de-prioritized.
When we have a reservation on a relay, and we lose connectivity to that relay, we should make an effort to reconnect to the same relay before searching for other relays to make a reservation on.
This is because making a new reservation will change our multiaddrs which impacts things like DHT provider records which will now be out of date just because our WiFi was a bit flaky for a time or we closed our laptop lid.
The text was updated successfully, but these errors were encountered: