Ask Alidade from Slack and Linear.
Engineering
Our read-only warehouse model, explained
Alidade connects with a read-only role, runs its queries in your warehouse and stores only aggregates.
Ben Achterberg
·
When you connect Alidade, we ask for one thing: a role that can read. It cannot create tables, write rows or change permissions. If you revoke it, we stop. This post explains what that means in practice.
What Alidade can see and store
Alidade sends SQL to your warehouse and reads the results. Row-level data is processed in memory and is not kept. What we store are aggregates: daily counts, rates and the segment totals needed to draw a chart or write an alert.
A typical workspace holds about 40 MB of aggregates for two years of history, even when the warehouse holds billions of events. Because the source stays in your warehouse, you can delete our copy at any time and rebuild it from there.
How to set it up
Create a role with SELECT on the schemas you want us to read, and give it its own warehouse or compute pool so our queries never compete with yours. On Snowflake, an XS warehouse with a 60 second auto-suspend is enough for most teams.
Every insight links to the query behind it, so your data team can read the exact SQL and run it themselves. If a query is slower than you expect, tell us and we will change it.
Setup takes about ten minutes. After that, there is nothing to maintain on your side.
