Running 10k multi-tenant GitHub Actions jobs a day with a small team (open source)

I Built This
1

I open-sourced Forge — a multi-tenant GitHub Actions runner platform that runs ~40 teams and ~10,000 CI jobs a day on AWS with near-zero ops — and wrote up the engineering trade-offs behind it.

It’s not a feature list. It’s the honest version: about a dozen deliberate decisions, each with a real cost named out loud — Calico to beat VPC IP exhaustion, a node pool per tenant for safe Docker builds, zero static credentials with a self-checking trust validator, immutable blue/green clusters, directory-as-config Terragrunt, and Renovate + dogfooded images so the platform tests its own changes on itself before merge.

Repo: GitHub - cisco-open/forge: ForgeMT is a secure, scalable GitHub Actions runner platform for ephemeral workloads. Designed for multi-tenant environments, it automates isolated runner provisioning on Kubernetes or EC2, with built-in OIDC, IAM, cost optimization, and deep observability. · GitHub
Write-up: No Silver Bullets: Engineering a Multi-Tenant CI Platform a Small Team Can Run - DEV Community

Happy to answer questions on any of it.

1 Like
2

Congrats on building this and already hitting 200 stars. I hope you’ll keep coming back and sharing milestones as your project grows.

3

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.