ScheduleMessagesOfDelayedAsync Limit

[PoteRii]

Hello,
If for some reason there is large number of records in db with Delayed or Queued status, all of the records are fetched into memory and this might create problems if you are running pods in K8s, because typically each pod does not have much memory. What do you think about adding a limit here? it can be configurable
https://github.com/dotnetcore/CAP/blob/master/src/DotNetCore.CAP.PostgreSql/IDataStorage.PostgreSql.cs#L240

[yang-xiaodong]

I'd prefer not to add this configuration, as the CAP scheduler is inherently lightweight by design.
How much data do you intend to load into memory?

[PoteRii]

It mostly depends on the feature and use-case, but limitless query from db into memory is always a bad idea...

[kvrivishvil1]

@yang-xiaodong we can create configuration with lot's of limit for example 100K or something like this but whoever will need to decrease it will decrease on his side.
We had issue few days ago and lot's of queued messages accumulated so everything went down for this service and we needed lots of apport to fix this

[tamazbagdavadze]

any progress regarding this issue? We have a lot of delayed messages and our services are consuming huge amount of ram at the moment