Logging in the age of Microservices and the Cloud

on YouTube ↗38m

" summary

Modern logging moves from SSH and tail to centralized structured logs stored as JSON, decoupling storage from presentation so raw fields remain queryable. Use an MDC or propagated request-id across services to trace a request end to end. Prefer managed collectors over DIY unless traffic exceeds extreme levels — CloudWatch comfortably handles 15,000 messages/sec per instance in practice. Standardize keys and values, tune retention to balance cost and usefulness, and capture context around errors rather than only severity.

AI-written summary · about

" chapters (15)

1. 0:00Intro
2. 1:30The good old days of logging
3. 3:20The cloud
4. 4:03Why are we logging
5. 6:03Capturing log information
6. 6:35Map diagnostic context
7. 10:12Querying logs
8. 12:52Cloud challenges
9. 15:25Autoscaling
10. 19:48Writing the logs
11. 23:10Log retention
12. 24:47Log messages
13. 28:12Distributed system
14. 28:35Trace request flow
15. 31:39Query logs

" related talks

• 39mObservable microservicesMaria Gomez, Taruj Mathur · Codemotion Madrid 2017
• 31mPast, Present and Future of MicroservicesSebastien Stormacq · Voxxed Days Microservices 2019
• 28mSurviving in a Microservices EnvironmentSteve Pember · DevOpsDays Boston 2018
• 24mMonitoring Microservices The Right WayDotan Horovits · stackconf online 2021
• 33mEffective observability in microservice architecturesLesley Cordero · LeadDev New York 2022
• 04mObservability in micro-service architecturesMatt Heath · dotGo 2014
• 31mHundreds of Microservices without breaking your APIsMichael Haberman · Conf42 Cloud Native 2021
• 52mSurviving in a Microservices TeamSteve Pember · Spring I/O 2017