they're what everything is generated off of

you can also submit a PR to update the rust client, or wait for us to

oh

i didn't realize there was one.

https://crates.io/crates/authzed is the only crate i found and it is in coming soon status

we may have un-published it

yea, that 404s

🤦‍♂️ it's private

if you decide to publish it, i'd use either rust-protobuf 3.0.alpha or prost to generate the client.

One of the folks in our community has one built with tonic

ah, forgot that prost doesn't handle the client gen. yea, that makes sense (tonic + prost are part of the same ecosystem).

prost 0.10 enabled the ability to run prost/tonic generators through protoc/buf

https://crates.io/crates/protoc-gen-prost should help

If I'm using managed SpiceDB from Authzed and I want to use the HTTP REST API, what domain do I use? I tried the swagger UI "try now" feature, but I just get back: request:

POST http://gateway-alpha.authzed.com/v1/schema/read
Authorization: <redacted, token provided in ui>

response:

{
  "code": 12,
  "message": "Method Not Allowed",
  "details": []
}

That's the correct endpoint but you'll want to use https

So I've changed to HTTPS and now I get a different error. request:

curl --location --request POST 'https://gateway-alpha.authzed.com/v1/schema/read' \
--header 'Authorization: <redacted>' \
--header 'Content-Type: application/json' \
--data-raw '{}'

response:

// HTTP Status Code 503
{
    "code": 14,
    "message": "connection error: desc = \"transport: Error while dialing dial tcp: missing address\"",
    "details": []
}

gateway

Also since you're using the swagger ui, you'll have to include "Bearer" with your token so "Bearer <your token" when you set the Authorization header

👋🏻 does

"rpc error: code = Unimplemented desc = unknown service authzed.api.v1.SchemaService"

sound familiar when doing

zed schema write

?

See a

V1SchemaServiceEnabled

flag

no that doesn't sound familiar

all of the services should get started automatically by the pkg/cmd machinery

I discovered that somehow I was setting "disable schema service v1" flag in my run configuration!

what is the difference between "dispatch" and "cluster dispatch" from the caching dispatcher perspective? https://github.com/authzed/spicedb/blob/94c69634c5d552fab34a9402e19973d7f68904da/pkg/cmd/serve.go#L59-L60

client and server side caches, e.g. if node A is dispatching to node B, there is a cache on both A and B with the subproblem and the computed answer

A's cache is much closer (nanoseconds) but is less likely to have the answer due to the way the consistent hashring works, B's cache is more likely to have the answer but is probably high microseconds to low milliseconds away

understood! 👍🏻 thanks for clarifying