--collector.otlp.enabled=true - --collector.otlp.grpc.host-port=4317

You just need these flags on Jaeger, basically

Jaeger itself is moving away from their own protocol, as OpenTelemetry APIs are standard

any suggestion?

alright I will have a look internally, but sounds feasible and also reasonable, given the status of the jaeger apis

thank you very much! 🙂

To deal with this internally we actually template our schema files, but in the future when authzed.com move to physical isolation between permissions systems, you won't need prefixes at all!

Do you have an example for template ?

any timeline you have in mind for physical isolation ?

not really, you can use something like moustache, jinja, go templates, etc

{{ ps_prefix }}/objectname

Gotcha

We have quite a few things left to implement, my guess would be Q1 next year

zed schema write

will also automatically rewrite prefixes if you're using it, but that means invoking the CLI vs the API

We need to do it via API

I figured 🙂

This should be fixed now if you pull again! Thanks for the report

pull the same

v1.14.0

image tag, right?

yep

Hi! I'm trying investigate an issue with SpiceDb being ran with docker-compose and I cannot exec shell on the SpiceDb container, apparently it doesn't have /bin/bash nor /bin/sh. So what is this image built on and what should I exec in order to attach to that running container?

I'm trying to bootstrap SpiceDb with 2 files: schema.yaml and relationships.yaml. Can you pls send me a link where I can read how to format relationships file properly, because right now it seems SpiceDb doesn't like my file and fails with "go panic" on bootstrap

the file format is the same produced by the Playground

its not formally documented but it can be found here as examples: https://github.com/authzed/spicedb/blob/main/internal/services/integrationtesting/testconfigs/quay.yaml

its a single file

I see that --datastore-bootstrap-files allows specifying 2 files, or it doesn't? When I specify only one file (contains both schema and relationships) everything is alright, but once I separate them and bootstrap as

--datastore-bootstrap-files="/etc/spicedb-relationships.yaml","/etc/schema.yaml"

it crashes with "go panic: memory" exception

I have a related question: is there a script for seeding SpiceDB? whenever i rebuild the DB, i start with a head migration then write my schema and finally execute a batch like:

# make user from accounts.id = 2 the owner of 
# the organization from organizations.id = 1
zed relationship create organization:1 owner user:2
# make events with events.id 1 - 27 belong 
# to the organization from organizations.id = 1
zed relationship create event:1 organization organization:1
zed relationship create event:2 organization organization:1

can i persist this somehow in a config of sorts and seed everything at once?

the question is: does --datastore-bootstrap-files really support many files?

further, since SpiceDB data is not always compatible eg i understood that v1.13.0 -> v1.14.0 upgrade had to be undergo deleting old data and starting from a head migration then reseeding relatiohships, is there a migration tool - ready or planned for persisting SpiceDB data and rebuilding SpiceDB again from such data

yes, but I wouldn't recommend it unless you only put relationships in the second, third, etc

much like MS SQL Server 25 years ago knew how to export its data into DDL

a planned utility is fine with me, too. just trying to understand what is the plan here...

this is not correct; the migrations are built in