Embracing Observability in Distributed Systems

preview_player
Показать описание
InfoQ Dev Summit Boston, a two-day conference of actionable advice from senior software developers hosted by InfoQ, will take place on June 24-25, 2024 Boston, Massachusetts.

Deep-dive into 20+ talks from senior software developers over 2 days with parallel breakout sessions. Clarify your immediate dev priorities and get practical advice to make development decisions easier and less risky.

----------------------------------------------------------------------------------------------------------------

Michael Hausenblas discusses good practices and current developments around CNCF open source projects and specifications including OpenTelemetry and FluentBit.

#DistributedSystems #Performance #Monitoring
Рекомендации по теме
Комментарии
Автор

Really good summary of the current ecosystem for obserrvability system and new architectures.

javisartdesign
Автор

01:43 Microservice characteristics
Pros: velocity, polyglot system, partial high availability
Cons: Distributed system, complexity, observability

04:13 Observability challenges
How to keep track of time and location?
What is the right retention period?
Return on investment
Different signals relevant to different roles

06:01 Observability end-to-end
Sources -> telemetry -> destinations

07:48 Dimensions based on morphological analysis
[Analytics, telemetry, languages, infra, compute unit, compute engine]

10:41 Signals
Logs: Discrete events, usually timestamped, can be structured
Metrics: Regularly sampled data points (timestamped), numeric value, dimension/labels
Traces: Request path across a number of services

12:11 Observability at CNCF
Graduated: Prometheus, Jaeger, Fluentd
Incubating: Cortex, Thanos
Sandbox: OpenTelementry, OpenMetrics, Chaos Mesh, Litmus

13:52 CNCF end-user technology radar
Observability, September 2020
Assess: Thanos, OpenTelementry, Kiali
Trial: Splunk, Sentry, Cloudwatch, Lightstep, Statsd, Jaeger
Adopt: Elastic, Prometheus, OpenMetrics, Datadog, Grafana

14:17 Routing logs with Fluent Bit
15:29 Prometheus & Grafana for metrics
Service mesh -> Prometheus > Grafana

16:54 Distributed tracing with X-Ray

17:52 OpenTelemetry

19:33 Summary
Freedom of choice: use the best of class for a use case
Rely on open standards such as OpenTelemetry
Build portable observability systems

21:18 Q&A

ErsinErtan