You’ve got AWS infrastructure humming along nicely, then someone says, “Can we run that on Cloud Run?” The room goes quiet. The AWS CDK team looks suspiciously at the Google Cloud folks, and suddenly the word “multi-cloud” starts flying around like confetti. That’s when you know it’s time to figure out how AWS CDK and Cloud Run play together.
AWS CDK (Cloud Development Kit) is the IaC framework that lets you define AWS resources in real languages—TypeScript, Python, the tools you already use. Cloud Run, on the other hand, is Google Cloud’s fully managed platform for running containers on demand. Each excels in its own ecosystem. But if your organization runs workloads in both, bridging them can feel like translating between dialects of the same language.
The goal is clear: deploy on AWS and Cloud Run with one repeatable, auditable process. AWS CDK can manage the AWS side—VPCs, IAM, Secrets Manager—while Cloud Run hosts microservices that don’t need to live inside AWS. You connect them through explicit permissions, service accounts, and identity federation. The logic is simple: let AWS handle stateful guts and let Cloud Run scale the on-demand compute.
To integrate, start by establishing identity boundaries. Use OIDC federation so that Cloud Run’s service identity can assume an AWS role. This avoids static keys and satisfies compliance policies like SOC 2’s “no long-lived credentials.” Then, automate the deployment sequence using CDK pipelines that trigger Cloud Run builds. The pipeline defines infra once, and each update redeploys code in both environments without human finger-pointing.
Common hiccups include mismatched IAM policies and timing issues when tokens expire mid-deploy. Fix that by granting minimal roles on each side and letting your pipeline manage rotation. Remember, you’re unifying clouds, not inviting chaos.