Quick question though: how would you handle open websockets on pods that are being scaled down? It seems that without any additional handling, those active connections will be cut prematurely, right?
From practice, they will get disconnected and retry to connect again. They eventually get connected/reallocated to a new pod. It happens almost instantly!
Sure, if the client was build with reconnect logic, then the client will reconnect.
So it seems like something to keep in mind when adding HPA to websockets: make sure the websocket clients can gracefully handle connection drops at any time 👍
ActionCable, Socket.io and WebSocket library do reconnect automatically when connection is dropped. But as you mentioned you should always think about such scenarios in planning a reliable software
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
Nice article!
Quick question though: how would you handle open websockets on pods that are being scaled down? It seems that without any additional handling, those active connections will be cut prematurely, right?
From practice, they will get disconnected and retry to connect again. They eventually get connected/reallocated to a new pod. It happens almost instantly!
Sure, if the client was build with reconnect logic, then the client will reconnect.
So it seems like something to keep in mind when adding HPA to websockets: make sure the websocket clients can gracefully handle connection drops at any time 👍
ActionCable, Socket.io and WebSocket library do reconnect automatically when connection is dropped. But as you mentioned you should always think about such scenarios in planning a reliable software