← Integration Engineering
Integration Engineering

Where Does the Data Go? Regions, Firewalls and Self-Hosting

21 Oct 2026·3 min read

"Our data must stay in the EU." If your integration cannot promise that, some deals end right there. For many businesses, where data is processed matters as much as what the integration does.

Regions and self-hosting

  • Regions: a region is a group of data centers in one area, like the EU or the US. Let a customer choose where their data lives. Be honest if logs or backups leave that region.
  • Self-hosting: for the strictest cases, like banks, run the integration inside the customer’s own environment. It costs both sides work, so offer it when truly needed.

Firewalls and IP restrictions

Some customer systems only accept traffic from known addresses, a list called an allowlist. So send your calls from a few fixed IP addresses, publish that list, and warn customers before it changes. GitHub publishes a list like this for its webhooks.

curl -s https://api.github.com/meta | grep -o '"hooks":\[[^]]*\]'
I ran this while writing. It printed a short "hooks" list of address ranges. GitHub’s docs say these change from time to time, so update your allowlist regularly.

The other direction works too. Let customers limit their API keys to their own IP ranges. Then a leaked key does not work from anywhere else.

Do this in 60 secondsRun the GitHub command above and look at the "hooks" list. That is what "publish your fixed IPs" looks like: a public list a firewall team can copy and check later.
Today’s AI angleWhen an AI agent calls tools, the data also goes to the model. So "where does the data go" now includes the model provider and its region. Put it on the same list.
Where Does the Data Go? Regions, Firewalls and Self-Hosting · Ali Huzaifa