How MS3 Supports 10,000+ Transactions Per Second with Kuma Service Mesh Across VMs a...
summary
MS3’s Tavros platform uses Kuma (Envoy data plane) as a language-agnostic service mesh to deliver 10,000+ transactions per second across VMs and Kubernetes by auto-injecting policies per namespace, auto-enrolling workloads in mutual TLS, and shipping traces to Jaeger + Elasticsearch for correlated logs and root-cause analysis. Kong handles ingress, Kuma enforces zero-trust via declarative policies (Open Policy Agent integration), and integrations can still augment traces; claim is that this runs at scale with no measurable app performance impact.
AI-written summary · about
chapters (16)
- 0:00Intro
- 0:55The Essence of Software
- 2:43Requirements Refining and Rapid Prototyping
- 4:03Inspection and Adapt
- 5:05Buy vs Build
- 6:02Domain Driven Design
- 6:25Core vs Generic Domains
- 14:14Tavros overview
- 15:01observability
- 15:37applicationspecific tracing
- 16:54Kuma
- 17:56Demo
- 20:01Application Level Data
- 21:01Log Aggregation
- 21:39Mutual TLS
- 22:52Conclusion
related talks
12mService Mesh In Action With Kuma: Security, Routing, and Observability
56mObservability in Action for your Service Mesh
11mWelcome & Opening Remarks
53mDo you need a Service Mesh?
07mGetting the most out of service mesh
20mKeynote: Service Meshes and Observability
24mObservability for Istio Service Mesh
40mBuilding a Scalable Observability Platform for Edge Compute Nodes