Travis Raines
Travis Raines
> Could we just choose the right listener in parentRef of the HTTPRoute? Yes--reviewing again while less tired, I glazed over `sectionName` because it wasn't named `listener`, but that will...
We had a guide that chose a very bad example value (10). I've updated that to the default. We'd still want to know the math though, since our target count...
The field is still required despite the default because that's how it prevents subsequent updates from deleting it AFAIK. It's irrelevant on create actions, but relevant for updates. This is...
If you check the admin API for this route, does it show the configuration you'd expect (the route does have `snis` set to the `example1.local` hostname from the TCPIngress)? You'll...
I'd like to remove the old code path entirely (I don't think there's any compelling reason to keep it around), and 3.0 would be a good time to do so,...
It is indeed more complicated. With the HTTPRoute translator as-is, I think the answer here is that multi-service backends just simply aren't eligible for such consolidation. We could potentially change...
In https://github.com/Kong/kubernetes-ingress-controller/pull/3008#pullrequestreview-1141603475 and https://github.com/Kong/kubernetes-ingress-controller/pull/3008#issuecomment-1278638285 we've found that HTTPRoute rules are much more difficult to consolidate than Ingress rules, as they allow for more complex matching (headers, methods) and transformations (filters,...
I think that risk for licenses is sufficiently mitigated by exfiltration having minimal impact so long as we're not completely oblivious. The test licenses only let you run Kong Enterprise:...
Yeah, that's reasonable. My point was more that even in the worst-case scenario (we blindly mash the "approve run" button and it executes exfiltration code) the impact isn't all that...