The story
Built by a data engineer
who got tired of surprises
Not the good kind.
I've spent years working as a data engineer - building pipelines on BigQuery, writing dbt models, wiring Looker dashboards together, keeping Fivetran connectors alive. The stack grows, the tooling evolves fast, and at some point you realise: you can't keep up with it all.
The problem wasn't that I didn't care. It was that keeping up with 10–15 tools means checking 10–15 different changelog pages, GitHub release feeds, and documentation sites - on top of, you know, actually doing the job. So I didn't check. And sometimes a breaking change would land on a Wednesday, quietly buried in a release note, and I'd find out about it when a Friday deploy broke something in production.
I looked for something that could help. There are general-purpose changelog aggregators out there - tools that track SaaS products, APIs, developer tools broadly. But none of them focused on the data stack. They didn't know the difference between a Snowflake deprecation that requires immediate action and a dbt feature flag that's optional. They weren't written for data engineers. They weren't built around the tools I actually use every day.
So I built DataStack Radar.
"The goal was simple: give me one place where I can see what changed across my data stack this week, with enough context to know whether I need to act on it right now or if it can wait until Monday."
It monitors official changelogs for BigQuery, dbt, Looker, Snowflake, Databricks, Fivetran, Confluent, Airflow, Trino and more. Every new entry is classified as breaking, warning, or info using AI - so you can see at a glance whether something needs your attention today or not.
The free weekly digest already replaces a lot of manual checking. Pro adds same-day email alerts when a breaking change is detected - no more waiting until Monday. Max adds Slack alerts, webhooks, and API access for engineers who need changes wired directly into their tooling.
This is a side project built and maintained by one person. If it saves you even one broken Friday deploy, it's doing its job.