Aapo Talvensaari

That should be doable. PR welcome, or I/we (at Kong) can take a look, but cannot promise exactly when.

> This is one of those things where route entity being used not just for matching but also for manipulating requests creates a problem. Well, isn't it just preventing the...

Yes, `host_header` got added some 4 months ago, while `preserve_host` has been there from a day one, and yes they sure feel a bit conflicting.

I understand "the upstream is closer", but people have proposed us to add all the service fields to route entity so that they could adjust some service params by route....

I don't think it is weird for single upstream to answer: customer-a.app.com AND customer-b.app.com Aka *.app.com. But yes, the whole `strip_path` and `preserve_host` are things I don't like that much,...

This would probably fix it: ```moonscript -- this will hold both paths and route names to prevent them from being -- redeclared by paths lower in precedence filled_route_names = {}...

@battlebyte sorry for not much traffic here. I have been thinking this. Ultimately I would like this to be added to Kong PDK like: ```lua kong.service.request.clear_query_arg() ``` I have been...

Update: The Ada bindings look quite fine: https://github.com/bungle/lua-resty-ada And the Ada library is proposed to be added here: https://github.com/Kong/kong/pull/13120

@nboddu009, Could it be the `sudo` is not preserving your `ENV`?

A couple of comments: 1. I would love to see tests added where we have `2.0` as protocol version in incoming (http2/grpc) and `2.0` on upstream when doing `grpc_pass` (not...