site-logo

JAWS PANKRATION 2024

site-logo
HomeNewsTimetableCfPCommitteePromotionFollow UpPrivacy Policy

Cost-Effective SLO Error Budget Monitoring with Athena and CloudWatch

Lv400

Lv400

2024/8/24 23:20 (JST)
セッション情報

Ensuring service reliability is crucial, and monitoring SLO error budgets is key.

In this session, I will show a cost-effective way using AWS services. Discover how to aggregate ALB logs with Athena, design custom metrics in CloudWatch, and implement burn rate monitoring with composite alarms.

Additionally, I'll cover metrics publishing with Lambda, Slack notifications, and visual dashboards. Join this session to explore effective monitoring options while keeping costs low.

Takashi Iwamoto

Takashi Iwamoto

- AWS Community Builders -

セッションカテゴリ
Analysis
関連AWSサービス
Amazon Athena
Amazon CloudWatch
Application Load Balancer
AWS Lambda
AWS Chatbot
etc
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 presenter, Takashi Iwamoto from ENECHANGE, discusses a low-cost method for monitoring Service Level Objectives (SLOs) using Amazon Athena and CloudWatch. As an AWS Community Builder, he shares his experience in cloud support. The presentation focuses on monitoring SLOs using 8 metrics and 4 alarms, which can be combined into a single composite alarm. CloudWatch is highlighted as a cost-effective solution, with an estimated monthly cost of $3.3 for the US East region. The speaker uses Application Load Balancer (ALB) as an example but mentions that the method can be applied to other services. He demonstrates how to use Athena to query ALB access logs and generate metrics. The SQL query examples provided show how to count total events, bad events (such as 500 and 429 status codes), and critical user journey requests within a specified time window. To implement this monitoring system, the presenter suggests using AWS Lambda functions to execute Athena queries periodically and push the results to CloudWatch as metric data. For cost optimization, he mentions an alternative approach using RDS to store aggregated data for longer time windows. The presentation briefly touches on notification options using SNS topics and chat bots, as well as the possibility of creating dashboards using CloudWatch's visualization features. In conclusion, the speaker emphasizes that this method allows for low-cost SLO monitoring using CloudWatch and Athena, which can be applied to various AWS services and use cases. He encourages the use of EventBridge or similar services to trigger periodic query executions for continuous monitoring.

戻る

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