You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We have some unit tests for the plugin backend (health checks, resource endpoints, and streams). We don't have any for the vector stuff, but I think integration tests would make more sense here.
Ideally we would have some integration tests which ran the plugin inside Grafana with some combinations of config (e.g. OpenAI proxy configured with either Azure or OpenAI; vector service configured with qdrant or Grafana Vector API), and call the various plugin endpoints. Not sure how we'd handle auth with OpenAI or Azure there though.
We have some unit tests for the plugin backend (health checks, resource endpoints, and streams). We don't have any for the vector stuff, but I think integration tests would make more sense here.
Ideally we would have some integration tests which ran the plugin inside Grafana with some combinations of config (e.g. OpenAI proxy configured with either Azure or OpenAI; vector service configured with qdrant or Grafana Vector API), and call the various plugin endpoints. Not sure how we'd handle auth with OpenAI or Azure there though.
It would be nice to also test that the grafana-experimental package also works with each version of the LLM app plugin, but those tests may need to live in a separate repository, either https://github.com/grafana/grafana-experimental or possibly https://github.com/grafana/grafana-llmexamples-app (e.g. inside one or more example Grafana plugins, for full E2E tests).
The text was updated successfully, but these errors were encountered: