Snowflake · Workload Governance

Native attribution stops
at the warehouse.

Snowflake is the deep worked example of the APEX loop. Anavsan is the Workload Governance Platform for Snowflake and BigQuery. Powered by APEX, it traces waste (not spend) to a workload and owner, assigns the fix, enforces a decision through your review process, and proves closure from the data.

200+ Snowflake signals feed the Private Knowledge Graph. Credits, warehouses, queries, storage, and Cortex get an owner — then a decision, then proof.

Capabilities

Where Snowflake depth lives

Feature pages nest under this hub — not as peers of the APEX engine. Warehouses, storage, Cortex, and the Native App.

Tools

Start before you connect

Assessments and the Cortex simulator — Snowflake-specific, no login required to begin.

Language

Explore terms that are Snowflake mechanics

Not a second APEX explainer — the warehouse-level words teams already use.

The loop on Snowflake

Credits, warehouses, Cortex — then proof

A short recap in Snowflake vocabulary. The engine write-up lives on APEX.

01 · Trace

Waste, not spend

200+ Snowflake signals rank addressable waste — idle warehouses, spilling queries, Time Travel bloat, Cortex token burn — not the busiest credit line.

02 · Assign

A named owner

The Private Knowledge Graph resolves the workload and owner from tags, roles, and evidence. When ownership is not knowable, it says so — it does not guess.

03 · Enforce

Through your review

The exact warehouse, query, or Cortex change goes through your PR process. PKG is for assignment and closure — not AutoApply into production.

04 · Prove

From the data

The finding closes when warehouse credits, storage bytes, or Cortex tokens recover. Nobody self-reports a win.

FAQ

Snowflake hub questions

What this page covers, how we connect, and how the Native App differs from the full loop.

What does this Snowflake hub cover?
This hub is the map of Anavsan on Snowflake. Cost and credit governance, Storage Intelligence, AI and Cortex visibility, the Native App, assessments, the Cortex Cost Simulator, and Snowflake-specific language such as query spillage. The APEX engine itself lives on /apex.
Does Anavsan read Snowflake table contents?
No — metadata-only ACCOUNT_USAGE-class signals. Anavsan reads query fingerprints, warehouse usage, storage and cost metadata. It never reads, copies, or moves table contents.
How long does it take to connect Snowflake to Anavsan?
Most teams connect in about five minutes. A keypair-authenticated service user with metadata-only access is enough. First attributed burn is available on day one.
How is the full Anavsan platform different from the Snowflake Native App?
The Native App is monitoring inside Snowflake; the full platform closes the loop. Marketplace gives credit and performance visibility with no data leaving your account. The full platform adds the Private Knowledge Graph for assignment and closure, the Enforcement Desk, PR-gated decisions, and findings that close when the metric recovers. PKG is for assignment and closure — not AutoApply. Read the comparison →

Govern Snowflake waste
through to closure.

Start a 14-day trial or take the Accountability Gap Assessment. Same loop — Trace → Assign → Enforce → Prove — on the warehouses and Cortex workloads you already run.