I'm trying to figure out if Azure Integration Runtimes (AIR) that are locked to a specific region and not set to auto-resolve are supposed to use random IP addresses. I have a data flow pipeline that reaches out to a Snowflake instance, but it keeps failing due to IP addresses not matching the published IP ranges for the region where the AIR is located. According to Microsoft documentation, all AIRs in the same region should use the same IP address ranges. I found this [link](https://learn.microsoft.com/en-us/azure/data-factory/azure-integration-runtime-ip-addresses) that outlines this, and it mentions allowing traffic based on the published IP address lists for specific Azure regions. However, it seems to reference the default auto-resolve runtime, which isn't what I'm using. If this behavior is expected, and the IP address can only be guaranteed through a VNET, then I need to address this discrepancy in their documentation. Has anyone experienced this or have insights?
2 Answers
It sounds like you've done your homework! Microsoft does publish a document with JSON data showing all the egress IPs for Azure services sorted by region. If you want to ensure guaranteed traffic, setting up a self-hosted integration runtime with a VNet is usually the way to go.
If you create a brand new Azure integration runtime and select your specific region, you should see the IPs working correctly. It's important to ensure that the type is still managed by Azure, just not set to auto-resolve. That way, you're less likely to run into random IP issues.
Related Questions
How to Build a Custom GPT Journalist That Posts Directly to WordPress
Cloudflare Origin SSL Certificate Setup Guide
How To Effectively Monetize A Site With Ads