This may need to be a blog post, but users tend to think ingress-nginx is simple because they can go to town with annotations. In reality, the features that they want are actually really complex to expose safely. This is why the project struggled with CVEs.

“Simple for whom” is the question

Same at DFW. Maintainer Summit is gonna get weird tomorrow

Slight clarification - CNCF counts contributions as "somebody who made a review, comment, commit, created PR or issue.
Contribution is a review, comment, commit, issue or PR." So not exactly lines of code, but in the same vein

Lines of code according to all.devstats.cncf.io/d/5/companie...

Ack, I think there are some actionable steps we can take here. FWIW, this was one of the things we agonized over during the beta and GA planning: we definitely don’t want to give the impression that we are leaving sidecars behind. I’d love to continue to get your (and others’) feedback here!

I’ll do my best to look into the draining issue soon; I should have some time as things slow down during the US holidays.

Just to clarify though, the reason this hasn’t been fixed yet isn’t because of ambient. We’ve just had some maintainer churn recently and are trying to backfill the expertise

Yeah I’m with you; I’ve followed one thread in particular where you’ve been active, and I feel for folks running into those issues just looking for a script to run to get unblocked.

In the short term, I’ll make sure this gets brought up in the next WG meeting, and….

I think what I’m hearing you say (correct me if I’m wrong!) is that you’re concerned that you and the collective of users you represent are being de-prioritized in favor of ambient and you feel that the reason for that is because certain vendors are incentivized to emphasize ambient more

🤔 I think that’s true generally, but tbh I don’t think John or really anyone at Solo falls into this too much. For every proprietary thing they launch, I can usually find at least 3x the investment in OSS which is fairly rare, especially in a post-ZIRP U.S economy

Oh that wasn’t at all what I intended to imply! To your point, your issues, advocacy, etc. are based on what’s important to you right? Whether the contributions are code or not, they usually can be derived by that individuals incentives, whether end user or vendor

Ah in that case, I see this as less vendor tension and more roadmap tension because Ambient is similar to an Istio 2.0 in some ways. Sidecars and Ambient both exist under the Istio umbrella but their roadmaps are a bit different

Just to make sure I understand your point: are you saying you don’t think vendors of an OSS project should have enterprise or closed source features for that project? If not, then what makes Solo’s telemetry stuff not an enterprise feature?

I trying to be more nuanced than that. Look at the native sidecar api in Kubernetes for example. The reason it’s a restartPolicy vs something else is because of API backwards compatibility changes. With ambient, we can move a lot faster because there’s a lot less baggage

I think that’s certainly one way to look at it. The way that I frame it is that the community is a group of individuals who work on the project for ~whatever incentives. If native sidecars is a big priority for someone, it’ll get fixed sooner. Ambient is a priority by several individuals so it GA’d

I don’t think I’ve seen anyone who wants to contribute to Istio of all projects who wants to contribute for no other reason than fun. I think yes we as maintainers need to say no to some feature requests for stability’s sake, but IMO, we need to do more to empower users to extend Istio

Let’s agree to disagree. Vendor customers aren’t inherently superior to community users. Lower barriers and dev velocity aren’t mutually exclusive with stable or even enterprise. Look at Kubernetes for example

What I’m hearing/inferring from @karlstoney.com (please correct me if I’m wrong!) is that “if the vendors don’t do it, it doesn’t get done because they’ve hired/captured all of the expertise”. We need to do more to empower end users to make meaningful contributions in a timely manner

And a big part of doing that well is supporting end users (like yourself) so that they can potentially make these sorts of changes/fixes themselves. IMO, this is our biggest opportunity for improvement, and I’m excited to make some progress here!!