#215
B+
Rank #127
Synapse
Reference homeserver implementation for the Matrix decentralized chat protocol. Standout trait is first-class federation, letting independently-run servers interoperate in one chat network.
68.3
/ 100
⚙ Full Stack
Python (Twisted-based) homeserver backed by PostgreSQL, with optional worker processes for splitting federation, sync, and media handling across multiple containers under heavy load. Redis coordinates worker communication.
📈 Scaling Analysis
Synapse's worker-mode architecture genuinely allows horizontal scaling of sync/federation/media traffic across multiple processes sharing one Postgres, which large deployments (matrix.org itself) rely on. It remains notably resource-heavy per-user versus purpose-built chat backends, and Postgres tuning becomes critical well before most other apps in this batch.
🚀 Running on Nexlayer
Minimum viable deploy is a synapse pod plus postgres.pod:5432; larger deployments add worker pods and redis.pod:6379 for coordination, all pulled through mirror.gcr.io. Federation requires externally reachable TLS on 8448, which needs explicit domain/ingress configuration beyond typical internal-only pod wiring.