This message was deleted.

Hey @Sumesh Sukumaran 😉 Someone will get to you soon to answer all your questions.

Hey @Sumesh Sukumaran, when using OPAL scopes, opal needs to perform synchronous git operations on the file system. We use opal-worker to perform the sync git operations asynchronously. The OPAL_BROADCAST_URI is used for different purpose than the OPAL_REDIS_URL, the broadcast_uri is used to broadcast messages between opal servers, the redis_url is used as the broker for the opal worker celery . OPAL_SCOPES is used to enable the scopes feature About the other questions I will ask my team and get back to you.

@Omer Zuarets Thanks for taking the time.

Re which broadcast channel we recommend. It really doesn't matter. Internally we use Postgres Listen/Notify , but all options are well and are used in production by multiple companies

Hi guys, Can you tell me about these variables? I am trying to setup OPAL Scopes. OPAL_WORKER_TOKEN=token, OPAL_BASE_DIR=/opal, OPAL_REPO_WATCHER_ENABLED=0

Hey @Sumesh Sukumaran, sorry for the delay, OPAL_WORKER_TOKEN is the normal OPAL server token the worker is using to authenticate against the server. OPAL_BASE_DIR is the base directory on the filesystem where git repos will be cloned into. OPAL_REPO_WATCHER_ENABLED disables the single repo git tracking ( you don’t need it when using OPAL scopes )

@Omer Zuarets Can we configure webhook with OPAL scopes instead of polling?

Unfortunately this is currently not supported, the only way OPAL scopes work is polling from the repo.

OPAL scopes is technically still in early adoption, We plan to add support for webhooks and scopes in upcoming versions of OPAL. BTW we are always eager to accept code contributions to the project. If you'd to help speeding up this feature, let us know and we'll provide guidance. CC: @Asaf Cohen