search

Netflix / falcor-router

A Falcor JavaScript DataSource which creates a Virtual JSON Graph document on your app server.
http://netflix.github.io/falcor
Apache License 2.0
104 stars 46 forks source link

Repo activity #190

Open limscoder opened 8 years ago

limscoder commented 8 years ago

Seems like there isn't much commit/pr activity within this repo.

Is Netflix still planning on using and supporting falcor-router going forward?

Is there a plan/roadmap for new features or is this the end of the line?

trxcllnt commented 8 years ago

From what I understand, they're still figuring out staffing internally. In the meantime, I have forks with bug fixes, new features, and performance improvements in our forks @graphistry/falcor. Working on docs as we go, but things are generally the same.

jhamlet commented 8 years ago

+1

;-j Jerry jerry@hamletink.com 510.432.4317

Grammar and spelling (mistakes) brought to you by iPhone.

On Sep 21, 2016, at 17:46, Paul Taylor notifications@github.com wrote:

From what I understand, they're still figuring out staffing internally. In the meantime, I have forks with bug fixes, new features, and performance improvements in our forks @graphistry/falcor. Working on docs as we go, but things are generally the same.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

limscoder commented 8 years ago

Is Netflix using falcor-router in prod or are they using a different falcor backend implementation?

sdesai commented 8 years ago

Hey @limscoder - definitely not the end of the line.

As @trxcllnt mentions, we are trying to work out resourcing priorities against the Falcor library and Falcor Router for the next quarter though, and hope to share a roadmap for both in the next month.

Netflix is also working on re-architecting our data access layer which will allow us to leverage the JS falcor-router (some of the work we have on the roadmap for the JS Falcor Router aligns around goals for error handling, logging and metrics in this regard). Currently we use a Groovy/Java Falcor router in production.

limscoder commented 8 years ago

@sdesai I think we've come to the same conclusions outlined in the talk you linked, and we're looking for something that's simple and scalable like falcor-router to handle routing between multiple backend microservices. However, we are bit worried about adopting something without a lot of commit volume and PR volume. Would love to see a roadmap.

sdesai commented 8 years ago

@limscoder Understandable.

On the falcor-router front, the main things we have lined up over the next quarter (dependent on resourcing) are:

  1. Error visibility and logging

    Errors are encapsulated in the JSONGraph, and returned as 200s, which doesn't lend itself to visibility into errors in production and the paths/routes they're associated with. Most production systems will have request/response level visibility (400s, 500s, etc.) but since these are 200s, we need to augment with path/route level error visibility.

  2. Performance metrics for paths/routes

    Same thing for performance metrics - most deployments have request/response level performance visibility, but not path/route level performance visibility

  3. Migration to RxJS5 for performance and visibility/debug benefits

  4. Comprehensive (client-to-server) error flow design doc

    We don't have a comprehensive Error document, highlighting how Errors flow through the falcor-router, through middleware, down to the client model, through the errorSelector and into cache. We wanted to take a stab at this overview doc, and use it to highlight any gaps in error handling hooks for users, we need to fill out.

But as mentioned, I'll put together an overall plan, with issues, where we can both discuss these in more detail, and anything which we may want to add on.

eddieajau commented 7 years ago

@sdesai, @jhusain, @blesh would it be possible to organise some community hangouts this year? Our company is in a position to help but there's no point in going off on a tangent that is not mutually beneficial.

Personally I'd love to see a workshop or four on router internals to help us understand what's going on under the hood, crafting a roadmap for 2017, garner feedback on pain points using Falcor outside of Netflix's use-cases, and then work out how we in the community that are investing in Falcor can chip in.

Thoughts?

Thanks in advance.

sdesai commented 7 years ago

@eddieajau - Can I reach out to chat? I think all of these (hangouts, workshops) are good ideas in general. We're still focused on some timebound internal priorities currently as far as engineers go, but we have a fairly close roadmap (I just need to spend some cycles to verify the rationale and fill out details/context around some of them) which I'm happy to discuss with you and get context on what you're currently working on, until we get a chance to refine and publish it.

eddieajau commented 7 years ago

@sdesai sure. Drop me an email/gchat on mamboblue at gmail dot com