Snowflake cost monitoring

Catch Snowflake credit burn before the invoice does

CostMon reads Snowflake's own organization usage views and turns daily credit consumption into one normalized view broken down by service type, usage type, and account, so the warehouse-compute line climbing from repeated resumes, or from warehouses sized bigger than their workloads, is visible while it's still cheap to fix.

Start free See how it works

Snowflake organization usage views — live today

Snowflake's warehouse billing has two mechanics that are easy to get wrong and invisible until the invoice arrives: every resume from auto-suspend bills a minimum charge regardless of how short the query is, and each step up in warehouse size roughly doubles the credit rate. A dashboard that polls every couple of minutes with an aggressive auto-suspend timeout ends up paying the resume minimum over and over instead of running efficiently. And because Snowflake's own usage views report credits at the org level with a real reporting lag, nobody notices the pattern until it's already a month of wasted resumes.

Cost anatomy

What actually drives your Snowflake bill

60-second minimum per resume: Every time a warehouse resumes from auto-suspend, Snowflake bills a minimum of 60 seconds of compute even if the query that woke it finishes in a few seconds.
Per-second billing after that: Past the first minute, billing runs per-second until the warehouse suspends again. So the resume minimum, not the query itself, is usually what a chatty workload is paying for.
Warehouse size roughly doubles cost: Each step up in warehouse size roughly doubles the credits billed per hour. Right-sizing, not just auto-suspend, is the other lever that moves the bill.
Auto-suspend timeout, tuned wrong: An auto-suspend set too aggressively against a bursty, frequently-polling workload (a BI dashboard refreshing every couple of minutes) forces repeated resumes, each eating its own 60-second minimum. The "savings" setting can end up costing more than a longer timeout would.
Serverless features, separate meters: Snowpipe, serverless tasks, Search Optimization, and Automatic Clustering run on Snowflake-managed compute and bill on their own credit meters, distinct from and additional to warehouse compute.

Why CostMon

Built for Snowflake spend, from day one

Daily, not a lagging org rollup

CostMon syncs Snowflake's organization usage views on a schedule with a trailing correction window for late-arriving data, so credit burn is visible daily instead of only reconciling weeks later.

Warehouse compute, split from the serverless meters

Normalized breakdowns by service type, usage type, account, and region separate warehouse-compute credits from the serverless meters that bill alongside them, so a climbing resume-and-sizing pattern is visible early, and you know which account and meter to investigate in Snowsight.

One number for finance and engineering

Stop reconciling Snowflake's usage views against a spreadsheet. Everyone looks at the same normalized credit spend, broken down the same way, next to the rest of your stack.

Read-only, cost-scoped credentials

CostMon connects with a Programmatic Access Token scoped to Snowflake's built-in ORGANIZATION_USAGE_VIEWER role. It's read-only and never touches your warehouses, data, or query results.

60sminimum credit charge every time a warehouse resumes
2×credits per hour roughly double with each warehouse size step up
Dailysync cadence, not monthly

Getting started

Connect, sync, see — for Snowflake

01
Connect with a Programmatic Access Token
Add a PAT scoped to the built-in ORGANIZATION_USAGE_VIEWER role. CostMon never touches your warehouses, databases, or query results, only usage views.
02
CostMon syncs and normalizes
Daily credit spend is pulled from ORGANIZATION_USAGE.USAGE_IN_CURRENCY_DAILY, broken out by service type, usage type, account, and region, and reconciled into the same normalized model as every other provider you connect.
03
See the line before the invoice does
Open a daily, service- and account-level view of Snowflake spend and see warehouse-compute credits climbing weeks before they land on a bill. From there, go to Snowsight to pin down which warehouse.

FAQ

Common questions about the Snowflake connector

How does CostMon connect to Snowflake?

Through a Programmatic Access Token scoped to Snowflake's built-in ORGANIZATION_USAGE_VIEWER role, read-only and scoped to usage and billing views. CostMon never requests access to your databases, warehouses, or query results.

How often does Snowflake cost data sync?

On CostMon's regular background schedule, reading a trailing window to account for Snowflake's own reporting lag on organization usage views. Recent days get corrected in place as more complete data arrives, rather than staying wrong.

Can I see Snowflake spend broken down by individual warehouse?

Not by warehouse name. CostMon reports at the grain Snowflake's own organization usage views expose: service type, usage type, account, and region, per day. That's enough to see warehouse-compute credits climbing (and to separate them from the serverless meters), but attributing that to a specific warehouse means going to Snowsight's warehouse metering views. CostMon tells you early that something moved and which account it's in; Snowflake tells you which warehouse.

Does CostMon replace Snowflake's own usage views?

No. CostMon reads the same organization usage views and normalizes them alongside AWS, GCP, and every other provider you connect, so your Snowflake total sits in one daily cross-provider view instead of a separate login. Snowflake's own warehouse metering views stay the place to go for per-warehouse and per-query detail.

What does the Snowflake connector cost?

A flat monthly rate per plan, not a percentage of your Snowflake spend or the savings CostMon helps you find. Free includes 2 connectors; Professional and Enterprise include unlimited connectors.

Connectors

Cloud, AI, and SaaS: all in one place

AWS, GCP, and Azure for cloud; Anthropic, OpenAI, Amazon Bedrock, Vertex AI / Gemini, and Helicone for AI and LLM spend; Snowflake and Databricks for the data cloud; Datadog, GitHub, and Vercel for the rest of the stack. Add CSV import for any tool without a native connector. That's 14 connectors, all normalized into the same unified view.

Connect a provider. See one number.

Connect your first provider and CostMon normalizes it alongside everything else you run. Your team gets one number everyone can check.

Start free See how to connect a provider