I think this has been an issue always, but recently it’s been more pronounced. Whenever I update the server to a newer version, the clients apps tend to get stuck. This manifests as either a continuous loop of trying to refresh the view or just a blank screen for the user. The situation of course clears with a clean refresh (refresh ignoring cache) but that’s just technical enough to be hard to remember for some users so they end up asking help as their “rocket is stuck again”.
While stuck, some of the clients end up doing some crazy constant loop, bombarding the server with requests and leading to a lot of logging and their user status indicator flickering between online and offline. This is not somewhat of a concern because it spams the server logs quite heavily, and I see a bunch of these: “Error, too many requests. Please slow down. You must wait 44 seconds before trying this endpoint again. [error-too-many-requests]”. I think the main culprits are the ones using electron clients from their desktops, but I’m not sure.
Is there a way to cleanly handle this? I guess it would be optimal if the apps automatically forced a clean refresh when they detect server version change. In the absence of that, is there an admin tool to force every client to re-login or some such? That might at least stop the request spam until they refresh manually.
- Version of Rocket.Chat Server: 6.4.4
- Operating System: Debian 11 (server), clients on linux & windows: electron, chrome, firefox
- Deployment Method: docker
- Number of Running Instances: 1
- DB Replicaset Oplog:
- NodeJS Version: v14.21.3
- MongoDB Version: 5.0.22 / wiredTiger (oplog Enabled)
- Proxy: nginx
- Firewalls involved: UFW on server