What is Drift Detection?
Definition
Drift detection works by taking a snapshot of a database's live schema and comparing it, object by object, against what the ChangeLog history says should exist there. Any mismatch, an extra column, a missing index, a table that shouldn't be there, gets flagged as drift, along with details about what's different. Running this comparison on a schedule, rather than only immediately before a deployment, catches drift closer to when it happens, while it's still easy to trace back to a specific cause. Waiting until deployment time to check means drift can accumulate for weeks or months before anyone notices, by which point figuring out what caused it is much harder.
Why drift detection matters
Database drift itself is the underlying problem, an environment that no longer matches its expected state, but drift is only costly if it goes unnoticed. Drift detection is what turns an invisible, slowly accumulating risk into something visible and actionable, ideally before it causes a failed deployment or shows up as a gap during an audit. It's also what makes the difference between drift being a minor cleanup task and drift being the root cause investigators spend days tracing after an outage.
How drift detection works
A drift detection check compares a database's actual schema, gathered by inspecting the live database, against the expected schema derived from the ChangeLog's history of applied ChangeSets. This comparison can run on a fixed schedule, on demand, or as a gate before a deployment proceeds, and typically reports differences at the object level, table by table and column by column, rather than as a single pass or fail result. Some drift is expected and harmless, like a temporary table created for a one-off investigation; effective drift detection lets a team distinguish that from drift that indicates a real gap in process. Results are typically reported per environment, so a team can see at a glance which environments are clean and which have accumulated the most unexplained change, rather than treating drift as a single organization-wide number.
How Liquibase helps
Liquibase Secure's drift detection compares an environment's live schema against its ChangeLog history and flags differences before they reach production, feeding results into Liquibase Change Intelligence alongside deployment activity and policy outcomes. Because it runs across every database platform Liquibase Secure manages, teams get one consistent view of drift instead of a different detection process, or none at all, per platform, and because detection runs on a schedule rather than only at deployment time, drift is usually caught within days of occurring rather than discovered months later during an unrelated investigation.
