Spotify's Honk agent rolls out 240 dataset migration PRs across downstream pipelines, saving an estimated 10 engineering weeks
Spotify had to move about 1,800 direct downstream data pipelines off two heavily used deprecated datasets within six months, across BigQuery Runner, dbt and Scio. It used Backstage lineage and Codesearch to find target repos, Fleetshift to orchestrate, and its Honk background coding agent to make the changes. Honk had no skills or MCP access at the time, so the prompt had to be complete; a context file Claude adapted from a human migration guide wasn't enough until field mappings were spelled out in tables. Scio pipelines were dropped as too varied. Most BigQuery Runner and dbt repos had no build-time tests, so owning teams tested manually before merging.
The numbers (as reported)
| Automated PRs | 240 |
|---|---|
| Scope | ~1,800 direct downstream pipelines across two deprecated datasets |
| Effort saved | An estimated 10 engineering weeks |
Facts
- Company / subject
- Spotify
- Tools
- Claude
- In-house tools
- Honk, Backstage, Fleetshift
- Codebase
- brownfield
- Work type
- Migration
- Stack
- BigQuery, dbt, Scio
- Published
- 2026-04-22
- Last checked
- 2026-11-02
What the source claims Practitioner (own blog/paper)
Claimed by the builder, writing about their own work. Not checked independently.
- 240 automated migration PRs rolled out, saving an estimated 10 engineering weeks 240 PRs; ~10 engineering weeks saved
“saving an estimated 10 engineering weeks in the process”
Caveats
Company writing about its own platform. Scio pipelines weren't migrated by the agent, and without tests in most repos the agent couldn't verify its changes. The post doesn't say how many of the 240 PRs merged. Publish date read from page metadata.
More stories by tool: Claude
Related topics: MigrationsBrownfield codebases