Websocket message size limitation in JupyterHub

JupyterHub Zero to JupyterHub on Kubernetes
,
1

I’ve been encountering kernel crashes while rendering graphs in JupyterHub, and I think this may be due to the WebSocket message size limitation, especially when dealing with larger datasets. Could you provide guidance on how to increase the WebSocket size limit for larger data? Using z2hj configuration.

2

If you’re using an ingress or load balancer you might need to increase the websocket/body size limit or timeout. For example, with Ingress Nginx you can do this by adding annotations to the ingress:
https://kubernetes.github.io/ingress-nginx/user-guide/nginx-configuration/annotations/

3

tornado / Jupyter Server also have their own websocket message size limits. Setting tornado websocket_max_message_size can be set to an integer in bytes (the default is 10MB).

# jupyter_server_config.py in user environment
# e.g. /etc/jupyter/jupyter_server_config.py in singleuser image
c.ServerApp.tornado_settings = {
  "websocket_max_message_size": int(20e6), # 20MB
}
1 Like