Announcing Impact Analysis in dbt Pull Requests

Every data team wants to move faster without breaking things. But maintaining speed with confidence is increasingly hard when data teams are responsible for a larger and larger surface area. Downstream breakage can be difficult to avoid.

For example, your team might manage hundreds or thousands of tables and dashboards and syncs. With so many dependencies, it’s cumbersome to make changes to dbt models and be aware of potential downstream impact on models or dashboards.

Today we’re excited to share that Metaplane integrates with GitHub to provide impact analysis on any dbt pull requests (PRs). Within your PR, you can know which downstream models and important business intelligence assets will be impacted by changes made to dbt models, share impact analysis reports with stakeholders, and see potential root causes for data incidents using dbt PR history.

See the blast radius of your changes before you merge code

One challenge that Metaplane users run into on a daily basis is understanding which downstream BI assets are impacted by a change in a dbt model. For example, if a data engineer renames or removes a column, they might only find out about broken Looker dashboards after the PR has been merged into prod and some downstream data user comes across a broken dashboard.

Rather than reactively receiving messages from frustrated stakeholders, Metaplane can now proactively let you know which downstream dashboards may be impacted by changes to dbt models in the PR.

By parsing column-level lineage in the warehouse and to any BI tool, Metaplane can analyze changes made to your dbt models, then post a comment providing downstream impact analysis before you merge the PR. Armed with this information, data engineers can decide to remove downstream dependencies or communicate with teammates that renamed or deprecated columns shouldn’t be referenced.

Share impact analysis reports with teammates

Whenever the Metaplane PR bot posts a comment with downstream impact, an impact analysis report is available within the Metaplane web app. The PR comment can be shared with internal or external stakeholders that own the downstream dependencies and need to be made aware of upcoming changes.

Identify the root cause of data incidents

One of the first questions our users ask when a data incident occurs is “did we change any models recently?” Keeping track of everything that is changing across tools can be challenging, but it’s critical to have context about recent upstream changes so that data teams can quickly identify the root cause.

After adding the Metaplane PR bot, any change made to dbt models will be tracked. If a data incident is opened in Metaplane and impacts models that were recently changed, Metaplane will show any PR that were recently created that impacted these models

The beta Metaplane GitHub dbt PR bot is now available to every user. If you’re interested in joining the beta, message your account manager in Slack, send an Intercom message, or email team@metaplane.dev. Or if you're not a Metaplane user yet, get started here.