As your organization scales, Snowflake can transform from a powerful data engine into a significant and unpredictable operational expense. When the bill arrives, you’re left scrambling through query histories to understand what drove costs up. This reactive approach to cost management is a common challenge. Instead of using Snowflake less, use it smarter.

Imagine having a real-time, shareable view of your entire Snowflake environment: credit consumption, warehouse performance, and storage costs, all in one place. By connecting Snowflake’s rich usage data to a powerful BI tool like Looker, you can create automated dashboards that provide exactly this. This guide walks you through not just how to build these dashboards but how to establish a sustainable, automated monitoring system. We’ll cover secure connections, essential metrics, and the strategic choice between a DIY setup and a fully managed solution.

Manual reports create a bottleneck. A data analyst pulls usage statistics, formats them in a spreadsheet, and emails the report. By the time it reaches a decision-maker, the data is already hours or days old. This process is inefficient and fails to capture the dynamic nature of cloud data warehouse usage. Think of it as trying to navigate a highway by looking only in the rearview mirror.

Automated monitoring solves these problems and connects directly to business value:

• Real-Time Insights: Dashboards update automatically, giving you an immediate, accurate picture of consumption. This lets you spot cost anomalies or performance degradation the moment they happen, not at the end of the month.
• Proactive Cost Management: Instead of reacting to a high bill, set budgets and create alerts that notify you when spending exceeds a threshold. This empowers teams to take corrective action, such as optimizing a query or resizing a warehouse, before costs spiral out of control.
• Empowering Stakeholders: An interactive Looker dashboard is far more intuitive than a static spreadsheet. It allows department heads, finance teams, and engineers to self-serve, explore the data, and understand how their work impacts costs. This builds a culture of accountability and cost-consciousness across the organization.

Ultimately, automated monitoring transforms Snowflake governance from a painful, reactive chore into a strategic, proactive function that optimizes performance and protects your bottom line.

Before visualizing any data, establish a secure and reliable connection between Looker and Snowflake. This foundation ensures your data is both accessible and protected.

Prerequisite: Understanding Snowflake’s ACCOUNT_USAGE Views

Snowflake provides a wealth of metadata and historical usage data through a special read-only database schema. The most impactful data lives within Snowflake’s ACCOUNT_USAGE views. These contain everything from credit consumption and query history to storage costs and user activity, typically with a latency of 45-90 minutes. You don’t need any special connectors or agents; just grant your Looker user the appropriate permissions to query this schema.

Step 1: Configure Secure Authentication (The Modern Way)

Username and password authentication was the standard for a long time. However, this method is less secure and is being deprecated for more robust, modern approaches. We recommend setting up key-pair authentication from day one for any programmatic access, including your BI tools.

Using key-pair authentication, generate a public/private key pair. Assign the public key to your Snowflake user, while Looker uses the private key to authenticate securely without sending a password over the network.

Here’s a high-level overview of the steps:

• Generate the Key Pair: Use OpenSSL to generate an encrypted private key and a public key.

• Assign the Public Key to a Snowflake User: Copy the content of the rsa_key.pub file and assign it to your dedicated Looker user in Snowflake using an ALTER USER command.

• Store the Private Key Securely: The rsa_key.p8 file and its passphrase are your credentials. Store them securely where Looker can access them.

This setup creates a more secure and manageable connection, laying the groundwork for a reliable data pipeline.

Step 2: Establishing the Looker Connection

After configuring your Snowflake user, connecting it in Looker is straightforward:

• In Looker, navigate to Admin > Connections > Add Connection.
• Select Snowflake as the dialect.
• Fill in your Snowflake host and database details.
• Select Snowflake JWT under Authentication.
• Enter your Snowflake username in the Username field.
• Provide the contents of your private key file (rsa_key.p8) in the Private Key field and the passphrase in the Private Key Password field.
• Test the connection to ensure everything is configured correctly, then click Add Connection.

Once connected, you can begin modeling the ACCOUNT_USAGE data and building your dashboards.

At Stellans, we’ve implemented this for dozens of clients, identifying three core dashboards that provide immediate value for cost and performance monitoring.

Key Dashboard 1: Credit & Cost Monitoring

This critical dashboard supports financial governance by answering: “Where is my money going?”

• Metrics to include:
  • Credits Used Over Time: A daily or weekly view of credit consumption, broken down by warehouse. This helps identify trends and spikes.
  • Cost vs. Budget: Track your cumulative spend against a monthly or quarterly budget.
  • Most Expensive Queries: A table showing the top 10-20 queries by credit consumption over the last 7 days. This is invaluable for pinpointing inefficient code.
  • Usage by User/Role: Understand which teams or processes consume the most resources.

Key Dashboard 2: Warehouse Performance & Efficiency

This dashboard helps determine if your warehouses are correctly sized. An undersized warehouse leads to slow queries and frustrated users, while an oversized one wastes money.

• Metrics to include:
  • Query Load: Number of queries running on a warehouse over time. This helps understand peak usage hours.
  • Queued vs. Running Queries: A high number of queued queries indicates an undersized warehouse or a need to adjust scaling policies.
  • Cache Hit Ratio: A high ratio means Snowflake efficiently reuses results, saving credits. A low ratio signals opportunities for query optimization or data clustering.

Key Dashboard 3: Storage Cost Analysis

Compute costs usually dominate Snowflake bills, but storage is a slow-growing expense that can catch you by surprise. This dashboard gives visibility into your data footprint.

• Metrics to include:
  • Storage Costs Over Time: A monthly view of storage costs (in TB and dollars) to track growth.
  • Growth of Key Tables: Identify the largest and fastest-growing tables. Useful for spotting runaway data ingestion or opportunities for archiving.
  • Time Travel & Failsafe Storage: Visualize storage costs dedicated to Snowflake’s data protection features.

Building dashboards is only half the battle. To make them truly effective, push insights to stakeholders and alert them when action is required. Looker excels at this.

You can use Looker’s scheduler to automatically deliver key dashboards or specific reports to stakeholders via email or Slack daily or weekly. For example, a weekly digest of credit usage can be sent to every department head, keeping costs top-of-mind.

More powerfully, create alerts that trigger proactive notifications. For example, set an alert to notify you when daily credit usage for a specific warehouse exceeds a threshold. This prevents budget overruns. Select a tile on your dashboard, define the condition (e.g., value is above 100), and specify the notification channel. This turns your dashboard from a passive report into an active monitoring system. Check creating user-defined alerts in Looker for detailed steps.

The steps above provide a solid roadmap for a DIY approach. However, maintenance introduces hidden costs. The Looker Marketplace offers pre-built “Blocks” for Snowflake monitoring, and competitors like CloudZero or Chaos Genius offer broad cloud cost management platforms. While these seem like quick wins, they come with challenges.

• DIY & Marketplace Blocks: Building or using templates means you own maintenance. When Snowflake adds features or changes ACCOUNT_USAGE views, your LookML models break. You must troubleshoot, update code, and ensure data accuracy. This diverts analyst and developer time from high-impact work to routine maintenance.
• Broad Cost Platforms: Tools like CloudZero are powerful but often designed for multi-cloud infrastructure monitoring. They may lack the granular detail needed for deep Snowflake optimization and often come with a hefty price tag.

Stellans offers a different approach. We provide a plug-and-play, fully managed solution that removes the maintenance burden. Our Snowflake Monitoring Dashboard Solutions are not just templates; they are a service. We handle setup, maintain LookML models, and ensure dashboards stay up-to-date with Snowflake features and best practices. This guarantees reliable, accurate dashboards with no technical overhead for your team.

Automating Snowflake usage monitoring with Looker is a critical step toward cost governance and performance efficiency. Establish a secure connection, build dashboards around key metrics like credit usage and warehouse performance, and set up automated alerts. You can transform your relationship with Snowflake from reactive to proactive.

While the DIY approach is feasible, it adds long-term maintenance work. A managed solution offers the best ROI by freeing your team to focus on insights instead of dashboard maintenance. Whichever path you choose, take action and bring real-time visibility into your Snowflake environment. Doing so turns data into a well-oiled machine that drives business value, which is central to our end-to-end data engineering services.

Ready to get real-time visibility into your Snowflake usage without the hassle? Contact Stellans to see a demo of our automated monitoring dashboards.

How do I connect Looker to Snowflake usage data?
You connect Looker by creating a secure database connection using key-pair authentication or OAuth2. Then, model data from Snowflake’s ACCOUNT_USAGE views within Looker to build visualizations and dashboards.

How do you monitor Snowflake credit usage in Looker?
Build dashboards that visualize data from the WAREHOUSE_METERING_HISTORY view. Key metrics include credits consumed per warehouse over time, daily/monthly cost trends, and identifying credit-intensive queries.

Can Looker automate Snowflake cost alerts?
Yes, Looker can automate cost alerts. Configure alerts on dashboard tiles to trigger email or Slack notifications when a condition is met, such as daily credit usage exceeding budget thresholds.