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
At scale under load, some beacons may wait for the /status call before sending registrations or may simply fallback locally if it takes too much time ; relayers also tend to return errors at tail, this can result in more blocks being unnecessary built locally just because a relayer took more time. Being able to have the PBS status call immediately returns 200 without querying the relayers reduces the time taken in a hot path of beacons, this helps with attestations/block proposals.
I see two downsides:
logs can be noisy: if all relayers are down, commit-boost PBS will still be called, will try to reach out to relayers, before returning an error back to the beacon, which will fallback locally,
if a relayer is badly configured, it may take time to realize it, for this commit-boost could actually check them once at startup, without blocking the hot-path.
The text was updated successfully, but these errors were encountered:
At scale under load, some beacons may wait for the /status call before sending registrations or may simply fallback locally if it takes too much time ; relayers also tend to return errors at tail, this can result in more blocks being unnecessary built locally just because a relayer took more time. Being able to have the PBS status call immediately returns 200 without querying the relayers reduces the time taken in a hot path of beacons, this helps with attestations/block proposals.
I see two downsides:
The text was updated successfully, but these errors were encountered: