Hi, I have been using rasa-x a lot recently, and with my recent models, everytime I try to talk to the bot inside rasa-x it crashes.
It has no problem training the models, and i can make changes to the nlu data just fine, but when I talk to the bot using the models, it crashes almost immediately. Even just “hi” crashes it.
I don’t really know where to start with debugging this? The models work locally. I’m able to run rasa shell in my environment and test just fine.
Any idea where to start with this? I think I should be able to check the logs with kubectl logs, but I don’t know which pod would contain the pertinent logs?
I’m using Rasa-X 0.30.1, I tried to upgrade to 32.2, but it isn’t upgrading. I changed the version in my values.yml file, and even uninstalled, and re-installed the whole chart, but it’s stuck on 30.1.
Okay. Can you describe what you mean by it crashing? With regards to the pods, the rasa-x pod and the rasa-production pod are the relevant ones with respect to the Talk to your bot page.
Yeah, when I talk to the bot, any message causes it to immediately crash. As in, the message disappears along with the story on the right hand side. It as if I pressed the clear button and the conversation never happened.
Okay, i’ve seen this occasionally before – the message never appears? (I’ve seen a bug before where everything clears and then re-appears when the bot message comes back)… I’d definitely check the rasa-production logs for this. You may need to turn on debug mode in order to get more info: Customize Your Deployment
I think I may have resolved this? For some reason the Rasa-x chart wasn’t updating after helm repo update, and was stuck on 1.6.4. Removing the repo, and adding it again actually updated the chart to 1.6.10 which seems to have resolved the issue.