Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.mycelium-ai.co/llms.txt

Use this file to discover all available pages before exploring further.

Databricks uses a webhook receiver (paid runtime) via Databricks Webhooks (MLflow) and Databricks SQL Alerts.

​
Webhook receiver (paid runtime)

POST /webhooks/databricks/{tenant_id}
Signature: Databricks webhooks send an Authorization: Bearer header with a configurable token. The receiver validates the token against the per-tenant secret.

​
Setup

Option A — Databricks Webhooks (MLflow experiment events)
  1. Create a webhook via the Databricks CLI or REST API: 
    databricks mlflow webhooks create \
  --events MODEL_VERSION_CREATED,MODEL_VERSION_TRANSITIONED_STAGE \
  --http-url-spec '{"url":"https://runtime.myceliumai.co/webhooks/databricks/{tenant_id}","authorization":"Bearer {your_secret}"}'
  2. Add to the runtime env: DATABRICKS_WEBHOOK_TOKEN_{TENANT_ID}.
Option B — Databricks SQL Alerts
  1. Create a Databricks SQL Alert on any query.
  2. Set the notification destination to a webhook pointing to https://runtime.myceliumai.co/webhooks/databricks/{your_tenant_id} with the Bearer token.
Option C — Unity Catalog audit logs Configure Unity Catalog audit log streaming to a Delta table, then use a Databricks Job to call the Mycelium External Function (same pattern as Snowflake).

​
Events ingested

Databricks eventMemory categoryNotes
Model stage transitiondecisionModel promoted to staging or production
SQL Alert triggeredexceptionThreshold-crossing on monitored queries
Job run succeededfactPipeline run result and duration
Job run failedexceptionFailure with error details
Unity Catalog auditfactData access and governance events

​
Retry behavior

Failed deliveries land in the per-tenant dead-letter folder. Replay via: 
POST /admin/webhooks/replay/{tenant_id}/{event_id}