As organisations scale their GraphQL adoption beyond a single team, schema federation becomes essential and complex. This talk explores the practical realities of building and operating a federated supergraph, highlighting the distinct strengths of Apollo Federation v2, Netflix DGS Federation, and GraphQL Mesh. We'll compare how each approach tackles entity ownership and boundary design, schema composition and breaking-change detection in CI, cross-subgraph authentication and authorisation patterns, and observability across a distributed graph. Drawing on real-world war stories, we'll address what documentation omits: how each solution manages schema governance across autonomous teams, handles subgraph versioning, and approaches debugging production issues spanning multiple services. You'll leave with concrete patterns for scaling GraphQL federation and a clearer understanding of where each federation approach excels, so you can avoid turning your supergraph into a distributed monolith.