Monitoring Network Traffic in Kubernetes

on YouTube ↗14m

" summary

Datadog shows concrete tools and outcomes: using AWS App Mesh with Envoy sidecars gives a central point of observability, and a new Datadog Network Performance Monitoring (GA tomorrow) watches layer 3/4 flows resolved to pod and service names. A trace example grows to 233 spans over 15 services and 16 hosts; a demo reroute triggers errors caused by insufficient pods visible as high TCP retransmits. Also offered is an open sourced watermark autoscaler used to reduce thrashing on a 2,500‑node cluster.

AI-written summary · about

" chapters (7)

1. 0:00Introduction
2. 1:04How we got here
3. 2:16Containers
4. 5:04AWS Mesh
5. 7:28Demo
6. 11:36Wrapup
7. 12:57Outro

" related talks

• 08mNetwork Performance MonitoringMiranda Kapin · Dash 2019
• 44mKubernetes at Datadog ScaleAra Pulido · DevOps Barcelona Conference 2022
• 34mHow to Survive to High Demanding ServicesEndika Gandarias · ContainerCon @ OSS EU 2023
• 14mLeveraging Distributed Tracing in Kubernetes & Containerized EnvironmentsChinmay Gaikwad · {unscripted} 2021
• 34mHow to Survive to High Demanding ServicesEndika Gandarias · Open Source Summit Europe 2023
• 46mObservability in the real world: Improve systems w/iterative approachJunaid Ahmed · AWS re:Invent 2022
• 44mChasing the Rainbow: Towards Unified Service MetricsDaniel Schierbeck · Dash 2022
• 42mDistributed Monitoring: How to understand the ChaosGianluca Arbezzano · DevOps Conference 2018