Fiddler has a dedicated integration-health view under the MLOps pipeline health section where you can track the status of each platform connection. Turn on alerts for sync failures there so a broken Databricks or MLflow sync pages you instead of silently going stale. This sits alongside Fiddler's model-level alerts on performance, data drift, data integrity, and service metrics like traffic, latency, and errors. The pattern that works well is: integration-health alerts for pipeline breaks, plus drift and data-integrity alerts on the models themselves, so you catch both the plumbing and the model behavior.

Context: How do I know if my Fiddler integrations are healthy? I want an alert when model sync from Databricks breaks instead of finding out from stale dashboards.