Manage schema migrations at scale

How do you manage database schema migrations in a high-scale, production environment?

Using the provided migration metadata, write a query that reports the current production migration health for every service, including services without a production migration.

Output

1. One row per service with service_name, migration_version, latest_status, failed_step_count, total_step_count, and duration_seconds
2. Include services without a production run, using NULL migration details and zero step counts
3. Return the latest production run per service, ordered by service_name ascending

Output details

• Break ties between production runs by started_at descending, then run_id descending.
• Count failed steps where status equals failed, and count every step in total_step_count.
• Return duration_seconds as whole-number elapsed seconds, truncating fractional seconds.
• Set duration_seconds to NULL unless both started_at and completed_at are present.
• Preserve every service record, including records sharing the same service_name.