ESDB has this smart functionality that allows maintaining less subscriptions if all the event handlers are live. In Eventuous context it would work like this:
Each subscription (group of handlers) is independent and can start from its own checkpoint
When subscription groups are catching up, they keep running independently
When a group starts to run live it switches to a single catch-up subscription, which runs live. The catch-up subscription used by the group gets disposed.
When all the groups run live, there's only one physical catch-up subscription running
If a particular group starts lagging behind, it switches to its own catch-up subscription
As a result, when all the groups are running live, there's only one physical catch-up subscription to the database, so there's less traffic and less read pressure on the database.
ESDB has this smart functionality that allows maintaining less subscriptions if all the event handlers are live. In Eventuous context it would work like this:
As a result, when all the groups are running live, there's only one physical catch-up subscription to the database, so there's less traffic and less read pressure on the database.
From SyncLinear.com | EVE-38