At Alvin, we know exactly what’s happening in your data warehouse. That’s because Alvin’s core is built on metadata events — everything from queries running on tables, schema changes, to shifts in usage and cost.
These events form observations about your system: a stale dbt model, an inefficient cluster config, or an unused table quietly inflating your bill. Until now, Alvin has been great at surfacing these observations.
But knowing what’s happening is only half the battle. The real challenge? Taking action.
That’s why we’re excited to introduce Workflows: the missing link between observing problems and fixing them.
Why Workflows?
Let’s be honest: data warehouses are noisy, costly, and sometimes messy.
- Stale models, poor configs, and unused tables sit around, quietly increasing costs.
- Engineers and analysts waste hours chasing problems and performing repetitive maintenance tasks instead of solving real business challenges.
With Workflows, Alvin brings action into observability. It’s not just about identifying problems — it’s about creating workflows to resolve them efficiently or, even better, automatically.
How It Works
- Detect Events
Alvin automatically detects inefficiencies in your warehouse and logs them as events. - Trigger Workflows
Workflows are triggered based on a subset of events that you care about, like stale dbt models, bad clustering, or cost anomalies. - Automate Actions
Once a Workflow is triggered, you can automate one or more actions, including:
- Create contextualized issue: Transform raw events into contextualized issues: What happened? How much does it cost you? What can you do about it? Sort issues by impact (e.g. potential cost savings) to prioritize the biggest wins. You decide whether to dismiss an issue, add it to your backlog, or take action immediately, right from Alvin.
- Generate a PR in Github to delete a model
- Switch data warehouse pricing model
- Send notification to Slack
- More automated actions coming soon! - Track Resolution and Impact
Once resolved, Alvin helps you verify the outcome: Did costs go down? Was the problem resolved fully?
Example: Stale dbt Models
Let’s walk through a common case:
1. On your Alvin homepage, you’ll see “Top Potential Monthly Savings”–a list of workflows ready for your attention.
For example, here are 16 potential “stale dbt models” issues, with a $1100 monthly savings potential on compute and storage costs.
2. Click a Workflow to see all related issues. The list is sorted by highest impact by default, so you can quickly prioritize what matters most.
3. The issue page explains the observation in detail and recommended actions.
4. In this case, we might fix it instantly by generating a PR from Alvin. Click the button, and a PR will appear in your GitHub repo that deletes the model in question, and provides the SQL to drop the table in your data warehouse. Done!
This is how the config of this workflow might look like:
Save Time. Save Money. Optimize Your Data Warehouse.
Workflows turn warehouse chaos into clear, actionable steps:
- Save costs: Identify inefficiencies and fix them fast.
- Save time: Automate repetitive tasks and streamline resolution.
- Focus on real work: Let your data systems “just work,” so you can focus on bigger challenges.
What’s Next?
We’re leaning heavily into automation:
- More semi-automatic actions (like PR generation)
- Fully automated fixes (actions based on rules)
- Real-time optimizations (more on that soon!)
Get Started Today
Alvin comes with many ready-to-use workflows based on our expertise in common problems. Setting up custom workflows is simple and intuitive. Don’t forget to connect Alvin to Slack to receive updates about your workflows.
We can’t wait for you to try it for free. Let us know what you think!
