MAF Multi-Agent on Cloud Run

A2A agents behind a load balancer, autoscaled, streaming to observability

MAF Multi-Agent on Cloud Run A2A agents behind a load balancer, autoscaled, streaming to observability Cloud Run (autoscaling) Clients · web / API · Architecture component Clients web / API Load Balancer · HTTPS :443 · Architecture component · global Load Balancer HTTPS :443 global Agent A · MAF service · Cloud Run (autoscaling) · Cloud Run Agent A MAF service Cloud Run Agent B · MAF service · Cloud Run (autoscaling) · Cloud Run Agent B MAF service Cloud Run Autoscaler · scale to load · Cloud Run (autoscaling) Autoscaler scale to load Observability · traces + metrics · Architecture component · prod Observability traces + metrics prod HTTPS A2A autoscale traces traces Legend External Cloud Backend

Edge

  • • A global load balancer fronts all traffic
  • • Requests fan out to independent agent services

Agents

  • • Each MAF agent runs on Cloud Run
  • • Agents talk peer-to-peer over A2A
  • • Cloud Run autoscales instances to load

Observability

  • • Both agents emit traces and metrics
  • • One production dashboard sees the whole system

Architecture diagram • Built with Archify • Create yours ↗ • Hover to trace • R route • Click to focus • +/− zoom • M radar • [/] views • P play story • T theme • E export