SpiceDB is an open source, Google Zanzibar-inspired, database system for creating and managing security-critical application permissions.

SpiceDB

So, there is no way to improve performance (in term of using cache effectively) for nodejs client. Am I correct?

the nodejs client calls the standard API, which includes caching

I mean to to use advantages of dispatching
```
This makes it critical that requests get dispatched to nodes that will have the data cached (if such a node exists), especially on Lookup requests which have lots of overlapping subproblems.
```

when a Check request comes in, it is broken into subproblems and each is dispatched to the correct node

the original node that is handling the (external) API request waits for the result(s)

Ok, so, we just need to enable `dispatch` function on spicedb containers, correct?

enable the dispatching system and make sure it has the correct resolution address

if under Kub, use the kubernetes resolver

this is one of the reasons that if you're running under Kub, we recommend the operator

https://github.com/authzed/spicedb-operator ?

documentation: https://docs.authzed.com/spicedb/operator

Hey, how do you like scheduling office hours? I’d like to show a schema design and ask for some feedback. Is it better if I just join voice channel whenever it’s done, or is it better to ask for a timeslot?

Probably better to schedule it. You can also DM me if you want to schedule a 1:1 call.

Hi, I am a newbie and planning to use SpiceDB for my org. We have lot of products in my org and I want to have a centralized SpiceDB cluster serving all these products. Is it possible for each product to define their own objects, relationship and permissions not conflict with other teams? something like a namespace for each product and roles, permission and relationships defined under this namespace and lookup can happen with a namespace specified?

is the intention that there is crossover between those products?

I would like to do both. Some products can crossover, and for few I want it to be isolated

if they are going to crossover, you'll want to use something like prefixes to specify which types belong to which products

`definition myproduct/someresource { ... }`

you'll need a process to combine the schemas into a single one before calling WriteSchema

but that's a good idea anyway for validation purposes

any docs/pointers on how this can be done? I would like each product teams to define their own schemas and maintain it

right now its fairly ad hoc - most users with this use case just have a process to append the team-specific schema fragments together to produce the final one that goes into source control

we're working on a proposal to make this more well-defined

the combined schema is then validated in CI/CD

that will be great, looking forward to it

if you want fine grained control over who can check and write portions of the schema

we're developing a solution as part of the paid offering

feel free to schedule a call at https://authzed.com/contact if you'd like to discuss it (and your other questions)

is the grpc-preshared-key(s) the only method of authentication to spicedb? Are there any options for the HTTP APIs?

The preshared key also applies to http requests through the gateway

If i'm were to have dozens/hundreds of clients, what is the recommendation? Create a proxy in front of spicedb that does some other authentication method, or initialize spicedb with dozens/hundreds of preshared keys and dole them out to different clients?