h2oai / h2oai/wave

Add support for detecting idle apps

Open
#608 4 comments 0 reactions 0 assignees View on GitHub
feature server
Dominant language
Python
Stars
4.3k
Forks
371
Avg merge
26d 1h
Merged PRs (30d)
1

Description

#### Is your feature request related to a problem? Please describe

In H2O AI Cloud, we need a reliable way fo telling whether an app is idle or not.
This should include both traffic on the FE (someone is actively looking at the UI) and asynchronous processing.
While the former can be handled automatically in the waved server, the latter needs explicit API in the SDK for the app developer to report if there is any background processing happening or not (we could use proxies like looking at OS-level processes, but that is not ideal)
This idle information then needs to be exposed on a handler for the Cloud service to collect.

#### Describe the solution you'd like

Provide a meta handler (e.g., /_idle) reporting whether the apps registered with the server are idle (or which are idle and which aren't), for some reasonable definition of "idle".

#### Describe alternatives you've considered

We could detect this in the Appstore server when proxying traffic to waved, but that wouldn't cover the backend processing very well.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.