

Databricks
QuotaGuard puts two fixed static IPs on your Databricks IP access list and sits at the connectivity layer, so every call from your app reaches the workspace from an address the list already allows. No change to your workspace architecture and no infrastructure to run. Trusted by data and engineering teams connecting to cloud platforms since 2013.
- Two-Minute Setup: Add your QuotaGuard connection URL to your app, set your Databricks client's proxy settings to point at it, and register your two static IPs on an ALLOW list. A workspace admin adds them once with the IP Access Lists API, the CLI, or the console.
- Works Across the REST API and SQL Clients: The Python SDK and CLI honor the standard proxy environment variable, and the Databricks JDBC driver, version 3 and above, takes UseProxy, ProxyHost, and ProxyPort. dbt, Power BI, and Tableau sit on those drivers, so one configuration covers your whole stack.
- Multi-Platform Support: The same configuration works whether you host on Heroku, Render, Railway, Fly.io, AWS Lambda, Vercel, Netlify Functions, Kubernetes, or a direct VPS. Set the proxy in your platform's settings and the client does the rest.
- Production-Grade Reliability: A load-balanced pair of static IPs with health checks and automated failover. Both IPs go on the allow list, so scheduled jobs and live queries stay connected through deploys and restarts.
- Shield for Regulated Lakehouse Data: For PHI, cardholder data, or anything under HIPAA, PCI-DSS, or SOC 2 scope, QuotaGuard Shield uses SSL passthrough so QuotaGuard never decrypts the data flowing between your app and Databricks. Static is not positioned for regulated data.
IP access list note: A static IP is needed only when IP access lists are enabled. The feature requires the Enterprise tier on AWS or the Premium plan on Azure and Google Cloud, supports IPv4 only, and must be turned on before any allow list takes effect. When you enable it, also allowlist the public NAT IPs your compute plane uses, or you lock out your own clusters. Databricks blocks applying a list that would lock out the user setting it, but that guard covers only that user, not your compute plane.