Skip to content

langstream apps deploy should provide more beginner friendly help if it can't connect to annot connect to http://localhost:8090/api/applications/ #709

@Gagravarr

Description

@Gagravarr

If you look at the README.md files for all the examples in https://github.com/LangStream/langstream/tree/main/examples/ , you will see a 50/50 mixture of langstream apps deploy test and langstream docker run for running the examples.

As a beginner, if you're aiming to just use the docker local runner to get started: If you first try a langstream docker run example (which will work), and then without spotting try a langstream apps deploy example next, you'll get a rather unhelpful error on the 2nd one

langstream apps deploy test -app examples/applications/compute-hugging-face-embeddings -i examples/instances/kafka-kubernetes.yaml -s examples/secrets/secrets.yaml
packaging app: /data/langstream/langstream/ghrepos/LangStream/langstream/main/examples/applications/compute-hugging-face-embeddings
app packaged
deploying application: test (2 KB)
Cannot connect to http://localhost:8090/api/applications/default/test?dry-run=false

In general, it might make sense for langstream apps deploy to print some additional help if it can't connect to the langstream server

For the specific case of langstream apps deploy trying to contact the default http://localhost:8090/api/applications/ URL, to be beginner friendly, I think it should offer quite a bit of help. Perhaps something like

The langstream server on http://localhost:8090/api/applications/ could not be contacted

Please make sure that the langstream server is running, the hostname (localhost) and port (8090) are correct, and any required kubectl port forwarding is in place
(possible link to some docs?)

And then if the app is also named test (as per all the examples), maybe also add

You can run apps locally without needing a full langstream cluster with langstream docker run

To do that, replace change your command from
langstream apps deploy
Into
langstream docker run

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions