Anavsan APEX Brings Workload Governance for Snowflake to Microsoft Marketplace

APEX evaluates every Snowflake workload and recommends one action: Run, Optimize, Redirect to better compute or storage, or Review. Your team approves it.

BOSTON, MA, UNITED STATES, October 27, 2026 /EINPresswire.com/ — Anavsan, Inc. today announced that the Anavsan APEX Governance Agent is listed on Microsoft Marketplace as a SaaS offer with a free trial. APEX governs Snowflake at the level of the workload: the dbt model, Fivetran sync, Snowpark job, Dynamic Table, Cortex call or Looker dashboard that actually generates the cost. It reads how each workload ran and decides where it should run next, across compute and storage. Snowflake customers who buy through Microsoft can start the trial from the listing and connect an account in about five minutes.

Every workload APEX evaluates gets one of four recommendations. Run means it earns its compute and stays as it is. Optimize means the same requirement can be met with less compute. Redirect means the workload belongs on a different path, such as a smaller warehouse, a serverless task, query acceleration, or a different table format and storage tier. Review means a person needs to look at it before it becomes scheduled spend. Each recommendation comes with the reason, the owner and the exact change.

Placement is where most of the savings sit. On compute, APEX reads SQL query fingerprints and dbt model history and matches each workload to the warehouse size, serverless task, or query acceleration setting that fits its cost and runtime profile. On storage, teams can run a pipeline on Apache Iceberg in parallel with the Snowflake native version and compare cost and output side by side. A parity check confirms the results match before anyone cuts over, and existing queries and BI tools keep pointing at the same tables.

APEX decides the route. The customer’s team makes the move. Every change goes out as a pull request through the review process the team already uses, and nothing reaches production without sign-off. After the change ships, APEX measures the same workload again to confirm the cost fell and the performance requirement still holds. The finding closes only when it does.

“Snowflake accounts don’t grow one warehouse at a time. They grow one workload at a time, from new applications, more pipelines, AI-written SQL and now agents, all landing on the same account,” said Bhuvana Palaniappan, Founder and CEO at Anavsan. “Native tools show what those workloads consumed after the fact. APEX looks at each one, decides whether it should run as is, run smaller or run somewhere else, and hands that decision to the team that owns it. Listing on Microsoft Marketplace puts that in front of teams who already buy their data tools through Microsoft.”

Here is an illustrative example from anavsan.com. An hourly dbt model runs on an X-Large warehouse for about 1,540 credits a month, finishing in 8 minutes with no queuing or spilling. APEX checks the same workload on a Large warehouse first: 11 minutes, still no queue and no spill, against a 45-minute requirement. The recommendation is Optimize to Large. Data engineering approves it, and APEX measures the next runs to confirm the result held.

Ownership comes from evidence. APEX reads tags, roles and the org map, and when it cannot establish an owner it marks the workload “needs owner” instead of guessing. Recommendations land in one Enforcement Desk, ranked by addressable waste rather than total spend, so a busy workload that runs efficiently stays off the list. Policy guardrails and budget alerts flag drift before it reaches the invoice.

Storage gets the same treatment. APEX flags unqueried tables, duplicate copies left behind by CTAS and cloning, and Time Travel retention set longer than anyone reads back. Cortex and AI Services spend is attributed to the agent, model or pipeline that generated it, a part of the Snowflake bill that resource monitors do not cover today.

APEX works from metadata only: query fingerprints, usage and cost signals. Table contents never leave the customer’s Snowflake account, and APEX does not sit in front of the query.

Also on anavsan.com

The same four decisions apply on Google BigQuery, where APEX evaluates jobs, slots, reservations and scan waste. Ask APEX lets FinOps and engineering question a recommendation directly and get back the SQL, assumptions, citations and an as-of timestamp behind it. A Private Knowledge Graph keeps workload identity, ownership evidence and fix history for each customer’s estate. Snowflake customers who want monitoring inside their own account can deploy Anavsan as a Snowflake Native App.

Pricing is the same on Snowflake and BigQuery. Monitoring is free for one Snowflake account or one BigQuery project. Accountability costs $1,750 per month per account or project, and Enforcement is priced at 3 to 5 percent of governed spend.

Availability

The Anavsan APEX Governance Agent free trial is available now on Microsoft Marketplace. Product details, an interactive example and pricing are at www.anavsan.com. Teams can book a 30-minute walkthrough at http://cal.com/anavsan/30min.

About Anavsan

Anavsan, Inc. builds workload governance for Snowflake and BigQuery. Its APEX agent reads how each data workload runs, decides whether it should run as is, use less compute or move to a better compute or storage path, and sends that decision to the team that owns it for approval. The company is headquartered at 867 Boylston Street, Boston, Massachusetts. Learn more at www.anavsan.com.

Media contact
Arun Manivannan
Anavsan, Inc.
contactus@anavsan.com

Bhuvana Palaniappan
Anavsan
+1 978-618-4738
Visit us on social media:
LinkedIn
YouTube

Media gallery