Hi all, I'm wondering if it is currently supported to perform failover while respecting the call load distribution algorithm.
What makes me doubt this is possible is the following wording in the documentation:
The rest of the destination list is taken in order of the entries in set - anyhow, until a re-route to next destination happens, the load on each address can change.
From it it's not clear if ds_next_dst() will simplify use next entry from the AVP list or if it will attempt to intelligently choose next entry from the AVP list taking into the account updated call load of the remaining entries.