site-logo

JAWS PANKRATION 2024

site-logo
HomeNewsTimetableCfPCommitteePromotionFollow UpPrivacy Policy

How to achieve full-stack Observability with AWS

Lv200

Lv200

2024/8/25 02:00 (JST)
セッション情報

In modern distributed complex systems like microservice architectures, observability is essential for quickly identifying issues.

While understanding the concept of observability as proposed by the CNCF, I will suggest methods to achieve full-stack observability utilizing AWS services.

Takashi Kaga

Takashi Kaga

- AWS Community Builders -

- AWS User Community Leaders -

セッションカテゴリ
Web and mobile frontend
Management and governance
関連AWSサービス
Amazon CloudWatch (Metrics, Logs, RUM, Application Signals)
セッション資料
セッションアーカイブ
セッションサマリ（by Amazon Bedrock）
    The presentation discusses implementing full-stack observability in AWS. It begins by explaining the Cloud Native Computing Foundation (CNCF) definition of observability: the ability of humans or machines to observe, understand, and act on a system's state. The speaker differentiates between monitoring and observability, with monitoring focusing on when and what errors occur, while observability aims to understand why and how they happen. CNCF defines three primary telemetry data signals: metrics, logs, and traces. Additional important signals include profiles and dumps. The talk then shifts to Amazon CloudWatch for achieving full-stack observability. Full-stack observability is defined as the ability to monitor the entire service stack, from frontend to backend infrastructure, including user experience and security. The presentation outlines a simple web service structure and focuses on frontend and backend monitoring. For frontend monitoring, it recommends using: 1. Amazon CloudWatch RUM for collecting user behavior metrics and logs 2. Amazon CloudWatch Synthetics for endpoint monitoring 3. Amazon CloudFront for outputting metrics and real-time logs For backend monitoring, the speaker suggests: 1. Collecting metrics from various AWS services 2. Using CloudWatch Application Signals for trace data 3. Implementing AWS FireLens for log forwarding 4. Utilizing Amazon Aurora RDS for database logs The presentation concludes by emphasizing the importance of understanding CNCF concepts and telemetry data, and encourages using Amazon CloudWatch for collecting and analyzing telemetry data in AWS workloads.

戻る

©JAWS-UG (AWS User Group - Japan). All rights reserved.