3.4 KiB
title, type, source-type, category, tags, date-added, video-source, audio-source, status
| title | type | source-type | category | tags | date-added | video-source | audio-source | status | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CTP Topic 67 Cloud native observability using OpenTelemetry | cloud-learning | video | DevOps & SRE/04_EKS |
|
2026-04-14 | nas:///volume2/work/Public Cloud Learning Sessions/CTP _ Topic 67_ Cloud native observability using OpenTelemetry.mp4 | summarized (Gemini 摘要) |
CTP Topic 67 Cloud native observability using OpenTelemetry
Source: NAS /volume2/work/Public Cloud Learning Sessions/CTP _ Topic 67_ Cloud native observability using OpenTelemetry.mp4
Type: VIDEO | Category: 04_EKS
Status: 🟡 Awaiting Whisper transcription → Summary
摘要
Surav from AWS presented a session on observability for Amazon EKS, covering the need for observability, code instrumentation using open telemetry, defining pipelines, AWS Distro for Open Telemetry collector deployment patterns, and observability deployment options on EKS and ECS.
Observability is essential for managing complexity as systems evolve. Building observable applications is a developer responsibility. Key signals to collect include traces, metrics, and logs, enabling reactive and proactive troubleshooting. AWS offers native options like CloudWatch and X-Ray, alongside open-source solutions such as Yeager, Zipkin, Prometheus, and Grafana, either self-hosted or managed. The AWS Distro for Open Telemetry (ADOT) is a secure, production-ready solution with AWS-developed components, offering support for operational issues.
Open Telemetry provides a vendor-agnostic instrumentation library, simplifying code instrumentation. The Open Telemetry collector uses receivers, processors, and exporters to manage signals. Receivers collect signals, processors transform them, and exporters send them to destinations. A trace captures the processing time taken at individual layers in your application call stack. ADOT includes the AWS SIG V4 extension for seamless integration with AWS services. Collecting metrics from both application and infrastructure layers allows comprehensive application views, including business-level metrics, service maps from X-Ray traces, and application logs. Correlation IDs, like the X-ray trace ID, enable deep links to trace views from log events.
ADOT is a repackaged Open Telemetry collector with AWS-developed components. It offers receivers like Prometheus and X-ray, processors like batch and filter, and exporters like X-ray, CloudWatch, Prometheus, and EMF. In ECS deployments, the AWS ECS container metrics receiver collects infrastructure metrics, while the Prometheus remote write exporter sends metrics to Prometheus. The SIGV4 Auth extension is used for AWS API calls. ADOT can be deployed as a sidecar container or a separate task, with configurations for scraping targets and defining pipelines. Deployment patterns include sidecar, separate task, demon set, and high-availability replicas. The ADOT add-on for EKS simplifies deployment with an operator and Terraform module, including prebuilt Grafana dashboards. Costs depend on the destination service, such as metric storage for Prometheus or trace ingestion for X-ray. An observability workshop and best practices site offer further guidance.
关键概念
行动项
相关视频
配对视频笔记链接（生成后填入）
最后更新: 2026-04-14