The task
Scenario. You are moving the legacy_dwh schema to a new warehouse, and the fastest way to stall is to migrate in the wrong order: a table you called low-priority turns out to feed forty dashboards, or has no owner to answer questions. You want every asset scored up front so you move the high-traffic, well-owned tables first.
The ask
- All tables in the legacy_dwh schema
- Each one's downstream consumer count, owner, and certification
- A high, medium, or low migration priority written to custom metadata
Hero prompt
We're migrating all tables in the legacy_dwh schema to the new warehouse. For each table, show me how many downstream consumers it has, who owns it, and whether it's certified. Then classify each one as high, medium, or low migration priority and write that to the Migration Priority custom metadata field.
Step by step
-
Search for all assets in the source environment.
List all tables in the <legacy_dwh> schema.
-
Check downstream consumers, owner, and certification.
For each, trace its downstream consumers outward (expand as needed), and show its owner and certification status.
-
Score each by migration priority and write it to custom metadata.
Classify each as high, medium, or low priority and write it to <Migration Priority> custom metadata.
What you'll give it
Required context
- The source scope (legacy_dwh)
- The priority rubric (what makes a table high versus low)
- A Migration Priority custom metadata field, and permission to write it (if it doesn't exist yet, create it once first — a one-time setup step)
Optional context
- A usage window for "active" consumers, so dormant dependencies do not inflate priority
- A cap, if you want only the top tier scored this round
What Atlan creates
Outputs
- Every legacy_dwh table with its consumer count, owner, and certification
- A high, medium, or low priority per table, against your rubric
- The priority written to the Migration Priority custom metadata field
Summary. The migration gets a running order instead of a guess. Sort by the field, and the high-traffic, well-owned tables move first, with the orphans flagged before they surprise you.
Changed: the Migration Priority custom metadata field on each table. Left untouched: the tables, their lineage, and the data, all read to compute priority, not edited.
Follow-up prompts
Keep going once the first answer lands.
Give me the wave plan
Group the tables into three migration waves by priority and show the owner for each wave.
Surface the risks
List the high-priority tables that have no owner or are not certified; those need attention first.
Track progress
As tables get migrated, show what share of high-priority tables remain.
Tips & troubleshooting
Sort by consumer count. Migrate the high-traffic, well-owned tables first.
- Match the tool to the work. This scores a whole schema; use a more capable model so the rubric holds across every table, and review before it writes.
- Automate or chain it. Re-run it weekly through the migration, so the priority field reflects what is already moved and what remains.
- Ask what it changed. Add "list every table you wrote a priority to and the value." That is the record of the triage run.