HN 표시: 구둣주걱(Shoehorn) - 유지 관리 오버헤드가 없는 백스테이지

Self-hosted beta Launching Like Backstage, but opinionated and ready to run. Point it at your repos and clusters; 15 minutes later you've got a catalog, search, docs, and workflows. Shoehorn runs as part of your stack. No portal team needed just to keep it alive. Three steps, and you're running. Point Shoehorn at your GitHub org, clusters, and repos. About 15 minutes of setup, then it runs like any other service you operate. Shoehorn crawls your repos directly. For clusters, drop in our k8s-agent and it pushes workloads to Shoehorn, with no inbound access required. You end up with what's running, who owns it, what it depends on, and where the docs live, without writing any of it. Search across everything, see what depends on what, and know who to page when something breaks, without first asking around in Slack. The three that come up most in Slack. It's 2am, something's down, and nobody knows who owns the payment service. Auto-discovered service catalog with team ownership, dependencies, and contact info Instant search finds services, APIs, and owners in under 200ms You want to deprecate an API endpoint. How many things call it? Impact analysis at configurable depth (1â5 levels) Dependency graph shows blast radius before you make changes Every new service starts with copy-pasting. Someone always forgets the CI config. Forge molds: reusable workflows to scaffold repos and provision infra Approval workflows ensure the right people sign off first Terraform, CLI, REST API, or the UI. Pick whatever fits your workflow. Manage Shoehorn config in Git, the same way you manage everything else Script and automate from the terminal. Get, put, and pipe. Every UI action is also an HTTP call, so you can automate from CI or scripts A point-and-click interface for the parts you'd rather not script Version-controlled and repeatable. Pick your interface.