Get Health

The MCP server’s get_asset returns the same four status enums (asset_health, materialization_status, freshness_status, asset_checks_status) inside its health field, and is enough when you only need the status itself.

Params: asset_key (path segment list, e.g. ["my_prefix", "my_asset"]), deployment_name (both required).

The dg api asset get-health command additionally returns the evidence behind each status, none of which the MCP server exposes:

• health_metadata — failing run ID, failed/missing/total partition counts, failed/warning/total check counts, last materialized timestamp
• latest_materialization — timestamp, run ID, and partition
• freshness_info — current lag, minutes late, maximum lag, cron schedule
• checks_status — check status with its counts

Use the dg command whenever you need to explain why an asset is degraded, not just that it is.