I just had to increase the timeout duration for the Ruby client because even though every notification is sent asynchronously, if even one takes more than 5 seconds (the current timeout):
- The client will produce a timeout exception.
- The pending async notifications apparently get canceled. We just published something and I realized it wasn't being published because I couldn't find it on my personal account's instance. But got notified when I increased the timeout duration.
I think shared inboxes can be helpful in this case #32 but making sure the notifications keep being sent after the connection is dropped would be great. Do clients need to know the remote instances got notified or that the Social Inbox received it and is handling it? Our client is not doing anything with this information, it just assumes the notification was received and sent when the connection is closed correctly (with an OK response).
I just had to increase the timeout duration for the Ruby client because even though every notification is sent asynchronously, if even one takes more than 5 seconds (the current timeout):
I think shared inboxes can be helpful in this case #32 but making sure the notifications keep being sent after the connection is dropped would be great. Do clients need to know the remote instances got notified or that the Social Inbox received it and is handling it? Our client is not doing anything with this information, it just assumes the notification was received and sent when the connection is closed correctly (with an OK response).