500 Error from Production

[Deprecated] Rasa X Community Edition
1

I’ve followed the Deploy to Server setup (docker) for Rasa X and successfully created a model from imported training data. I’m getting a server error when using the Talk to your Bot page.

I run the version call against Rasa X and the output shows:

{
    "rasa": {
        "production": "1.1.7",
        "development": {
            "status": 500,
            "message": null
        },
        "worker": "1.1.7"
    },
    "rasa-x": "0.20.0"
}

Should I have a development instance of Rasa running after following the Deploy to Server instructions?

When I go to the Models screen, it allows me to activate a model but there’s no choice of production or development, probably because development is not running. I choose Make Active on the model.

Then, go to the Talk to your bot and enter a question to the production bot. The network tab of the browser shows I’m getting an error 500 against the following URL.

http://botserver/api/chat?environment=production

Any suggestions on next steps.

2

There are 2 versions of Rasa, 1 is CE (Community Edition) and one is Enterprise, I remember installing from server there was a step to be sure your using CE edition. I believe the choice for Prod vs Dev is only in the Enterprise edition, community just gives you 1 to use. So if you trained a model it should autoload your model as prod.

Something might be funny with your imported data too, can you look at your data in Rasa X? ie stories etc?

3

Thanks for the clarification. Yes, I’m using the CE version which only has the one instance.

I’ve reviewed my data and did some clean-up based on warnings on the Domain screen and an error that occurred during Training. My training now runs without errors and produces a model which I can activate.

1 Like
4

excellent!

I have also installed rasa with the regular command and -nohup the Rasa X on a digital ocean server, used MobaX to transfer files over. I am new to a lot of this, but found it easy to drag story files etc, over, I have built a few bots comprised of different sample bots, and easily updated all my files that way

5

Unfortunately, it’s still not working. I have a second CE running that works fine. Not sure yet what the difference in the two instances.

What size DO droplet are you using?

6

I was running an 8GB, 40/month I think? I forget

I build and destroy, I think its the min that they recommend