Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Serve] Add deployment graph cookbook #24524

Merged
merged 6 commits into from May 11, 2022
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
5 changes: 5 additions & 0 deletions doc/source/_toc.yml
Expand Up @@ -67,6 +67,11 @@ parts:
- file: serve/http-servehandle
- file: serve/ml-models
- file: serve/deployment-graph
sections:
- file: serve/deployment-graph/deployment-graph-e2e-tutorial
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

you know I'm against this "feature-first" approach to docs hierarchy, but I can't stop you guys from doing your thing.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We're referencing Ray dataset structure which i believe you contributed a lot to its format, open to changes :) iiuc you're suggesting we should bring that to first level user guide instead? I think that's better format too to avoid nesting docs, but meanwhile Deployment graph was just in alpha stage but i haven't found a structure that satisfies both yet. What do you think is more reasonable?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm working on Serve doc restructuring. This structure will be transient for now.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for clarifying, @simon-mo . We don't have to change anything for this PR, just wanted to chime in.

@jiaodong Generally, users are interested in their workloads and workflows, rather than specific features of a library (that they often don't know yet and won't click on). A user guide / cookbook for this is great, but I wouldn't put it top level. To me there shouldn't be a "Deployment Graph" doc there at all, as I find the structure a little confusing.

Screenshot 2022-05-10 at 18 25 46

Where should I go to learn about deploying my models first? Core API: Deployments, Calling Deployments, Serving ML Models, Deployment Graph, or Deploying Ray Serve? They all sound very similar and I don't want to read through 5 docs first to understand what I need, because I might only have 10 minutes. All relatively easy to streamline, I think. No worries there.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the inputs Max, it makes a lot of sense. As we chatted earlier in March, Serve team should re-structure our docs to better organize our content on this and it's happening now, please don't hesitate to step up and give us suggestions like this :)

For this PR in particular, Sihan is ramping on the context to make progress, but I will self-assign and ensure syncing with Simon to properly organize them as part of the docs restructure effort.

- file: serve/deployment-graph/deployment-graph-user-guides
sections:
- file: serve/deployment-graph/chain_nodes_same_class_different_args
- file: serve/deployment
- file: serve/monitoring
- file: serve/performance
Expand Down