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
{{ message }}
This repository has been archived by the owner on Aug 16, 2021. It is now read-only.
Currently, the algorithm for picking realms is the following:
Get status and calculate latency of all the realms available through DAO (some realms will have the same latency since they belong to the same server)
Select all the candidates that are healthy and are not full
Sort them using the following criteria:
If latency difference is greater than 1500 ms, use that as the main factor
If the latency difference is within 1500 ms, consider the amount of users, and use the score function:
The score function prioritizes realms that have users and gives greater scores to those realms that have close to 40% of their main users, and deprioritizes those that are over 80% full. See here: feat: New score function #604
If the score is the same, the realms of those catalyst (including all layers) with the least users are prioritized, in order to balance the load.
Pick the first one
We should adapt this considering that now we don't have layers and we don't have a clear maximum amount of users per catalysts (in practice, it would several tens of thousands)
The text was updated successfully, but these errors were encountered:
Currently, the algorithm for picking realms is the following:
We should adapt this considering that now we don't have layers and we don't have a clear maximum amount of users per catalysts (in practice, it would several tens of thousands)
The text was updated successfully, but these errors were encountered: