---
title: "Cloud Run"
description: "2 posts about Cloud Run, summarised, each linking to the original."
---

# Cloud Run
> 2 posts about Cloud Run, summarised, each linking to the original.

## Articles

### [How to securely connect internal Cloud Run MCP servers to Gemini Enterprise](https://yomu.fyi/post/how-to-securely-connect-internal-cloud-run-mcp-servers-to-gemini-enter.md)
- Company: [Google](https://yomu.fyi/company/google.md)
- Author: Neeraj Agrawal
- Published: Nov 20, 2026
- Type: Tutorial

Enterprise teams using Gemini Enterprise build custom MCP servers on Cloud Run to connect AI agents to internal APIs, databases, and workflows. Three challenges arise: internal-only ingress blocks public calls, Cloud Run IAM and user OAuth 2.0 must both be verified, and Shared VPCs cannot use wildcard \*.run.app DNS zones without breaking other services. The solution uses an Egress Agent Gateway with Private Service Connect to route traffic privately into the VPC, then to Cloud Run via Private Google Access or a global PSC endpoint. FQDN-scoped Private Cloud DNS resolves only the MCP server hostname to a private VIP, avoiding DNS collisions. Gemini Enterprise sends two authorization headers: an OIDC ID token for Cloud Run IAM and the user’s OAuth 2.0 access token for application-level validation. A companion GitHub repository provides a runnable Python MCP server with OAuth 2.0 middleware and step-by-step setup scripts.


### [Building Overexposed: understanding application exposure (2026) - Shopify](https://yomu.fyi/post/building-overexposed-understanding-application-exposure-2026-shopify.md)
- Company: [Shopify](https://yomu.fyi/company/shopify.md)
- Author: Simar Singh
- Published: Oct 19, 2026
- Type: Problem & solution

Shopify built Overexposed, a scanner that checks which applications have internet-facing endpoints and whether they allow anonymous access. It scans Kubernetes Ingresses on GKE and Cloud Run services by querying GCP’s Cloud Asset Inventory for deployed resources, then sending anonymous HTTP probes—HEAD requests first, up to three redirects, checking for authentication challenges. The scanner classifies each endpoint into six verdicts—exposed, expected, protected, not found, unreachable, or inconclusive—based on the probe response and an allowlist of approved public endpoints. Overexposed reduced inconclusive Ingress results by about 80% within weeks by refining classification rules and testing changes against historical scan data. The project revealed that reading configuration files is insufficient because they show intent, not enforcement; the true inventory must come from what the cloud actually runs, and each verdict includes a reason for investigation.
