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
org.atmosphere.websocket.maxTextMessageSize should be allowed configure per handler, global one make all handler use same buffer size for Session. this make memory usage can increase by unnecessary.
Such as there are several end point, one is for large message, the max could be 10M, but others are designed for small message less than 1k. but the global config make all session buffer size 10M, this make the service use more memory is required, sometime is message is busy, it lead to outofmemoryerror.
The text was updated successfully, but these errors were encountered:
@nwong4932 I don't understand exactly what you are asking for when you talk about "per handler". But if you have one application that uses significantly more memory than the others, you should probably create a separate web app (i.e., a separate servlet endpoint) for that memory intensive service. As the size limit value needs to be passed to the underlining websocket factory at its startup, I don't see any flexible way of using a different value within a single endpoint. Or do you have a concrete suggestion?
thanks.
Hi Elakito,
I can't separate the user, they are in same app, the big message only send occasionally. Per handler means per managedservice annotation class. I do see tomcat jsr365 endpoint has its message size for every endpoint.
Best
org.atmosphere.websocket.maxTextMessageSize should be allowed configure per handler, global one make all handler use same buffer size for Session. this make memory usage can increase by unnecessary.
Such as there are several end point, one is for large message, the max could be 10M, but others are designed for small message less than 1k. but the global config make all session buffer size 10M, this make the service use more memory is required, sometime is message is busy, it lead to outofmemoryerror.
The text was updated successfully, but these errors were encountered: