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
We have deployed our push service using a2 into production, but we are frequently observing ConnectionReset errors on http2 connections.
Our service is multi-tenant and the traffic is a steady stream of messages with idle periods and bursts in between.
Although we are working around these connection resets by retrying it. I was thinking if we could do something like sending periodic keep-alive messages over the open connection.
I opened a PR to send the PING every 1 hour as per the recommendation by the APNs docs. Nothing though that there may be other factors causing connection resets (e.g. APNs closing the connection or network problems) so you should still use retries.
Dear team,
We have deployed our push service using a2 into production, but we are frequently observing
ConnectionReset
errors on http2 connections.Our service is multi-tenant and the traffic is a steady stream of messages with idle periods and bursts in between.
Although we are working around these connection resets by retrying it. I was thinking if we could do something like sending periodic keep-alive messages over the open connection.
The pushy library already does something like this: jchambers/pushy#266
I read through the hyper documentation, and the default is to not send any keep-alives
The text was updated successfully, but these errors were encountered: