Goal: I’d like to know what percentage of requests served through the gateway and resolved through the DHT are served within the first 10mins of the CID/Provider Record being published.
Context: Routing tables are refreshed every 10 mins and after the first 10mins of the CID/Provider Record being published the multiaddress of the original publisher is not included in the record anymore - that's in order to cover the case where a peer has gone offline and back online and it's multiaddress is not the same anymore. So one way to figure out what I need is to count the number of provider records returned to the gateway and include a multiaddress (as compared to those that don’t include a multiaddress). In that second case, clients have to walk the DHT again to map the PeerID to the peer's latest multiaddress.
A small modification on a gateway's kubo node should be able to do the trick. It would be great to have that in the next 3 weeks or so (if possible), as it will influence some other design decisions.
Goal: I’d like to know what percentage of requests served through the gateway and resolved through the DHT are served within the first 10mins of the CID/Provider Record being published.
Context: Routing tables are refreshed every 10 mins and after the first 10mins of the CID/Provider Record being published the multiaddress of the original publisher is not included in the record anymore - that's in order to cover the case where a peer has gone offline and back online and it's multiaddress is not the same anymore. So one way to figure out what I need is to count the number of provider records returned to the gateway and include a multiaddress (as compared to those that don’t include a multiaddress). In that second case, clients have to walk the DHT again to map the PeerID to the peer's latest multiaddress.
A small modification on a gateway's kubo node should be able to do the trick. It would be great to have that in the next 3 weeks or so (if possible), as it will influence some other design decisions.