-
Hey folks, I've exhausted my resources here trying to figure out why I can't actually acquire a subscription to a channel, exclusively, in my production environment. I'm deploying on DigitalOcean, with a very simple nginx configuration - which, has, had some issues websockets - however, I don't believe the connection itself is being severed in any way, primarily, because the client is receiving ping messages back. It appears to me that, for some reason, RPC commands are just not hitting the rpc server in the Kubernetes environment vs run locally. JavascriptI am using the AnyCable JS client (previously, the Rails ActionCable client -- same result) Javascript CodeProduction (Failing)Local Development (Succeeding)NetworkIn the network tab, it's clear, I'm receiving ping's from the Production (Failing)Local Development (Succeeding)Rails CodeThere's nothing special happening on the backend. As far as I can tell (based on print out statements) the channel code is never actually being called in Rails Channel CodeRails Cable CodeSever LogsThese are the Rails server logs, just for a single attempt to connect to the Server. Production, Kubernetes instances (Failing)Instances are roughly a 1:1 with the suggested helm file document: https://docs.anycable.io/deployment/kubernetes Local Development (Succeeding)Procfile:
RedisI confirmed with (connecting to dev and prod redis instances):
that my messages are actually passing through the |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
Unfortunately, this was simply a manifestation of CloudFlares "websocket" option not being turned on I simply, did not, realize that this setting had been disabled! Thanks! |
Beta Was this translation helpful? Give feedback.
Unfortunately, this was simply a manifestation of CloudFlares "websocket" option not being turned on
I simply, did not, realize that this setting had been disabled! Thanks!