SCaLE 16x - Application Monitoring And Tracing In Kubernetes
SCaLE 16x - Application Monitoring And Tracing In Kubernetes Session Info: https://www.socallinuxexpo.org/scale/16x/presentations/application-monitoring-and-tracing-kubernetes-avoiding-microservice-hell Creating and deploying microservices is easy. The real problem is how to manage and support these services out in the wild and in production. What happens when these services stop working or worse yet when they are running but running slowly? Which service instance is the culprit? This session talks about how you can leverage metrics and tracing tools in order to give better visibility into the distributed nature of a microservice architecture in a Kubernetes environment. This presentation will discuss key concepts of metrics and tracing and highlight Open Source Projects available that address: 1) the value of instrumentation and how Prometheus can be used to monitor and measure a Kubernetes cluster, and 2) how Jaeger can provide visibility into your applications and microservices. Will provide a demo of Microservices leveraging Prometheus and Jaeger deployed to Kubernetes cluster.
SCaLE 16x - Application Monitoring And Tracing In Kubernetes
- 1. Application Monitoring and Tracing in Kubernetes: Avoiding Microservice Hell! David vonThenen @dvonthenen http://dvonthenen.com github.com/dvonthenen
- 2. Agenda • Why do we care? • Introduction to Metrics • Introduction to Tracing • Demo • Q&A • Free give away! Slim Power Bank/Chargers!
- 3. Why do we care?
- 4. Microservices Are Awesome! • Discrete Set of Functionality • Resilient / Tolerates Failure • Distributed / Highly Scalable • Technology Freedom • Autonomy of Dev Teams • Enables Continuous Delivery
- 5. Can Be Your Worst Nightmare! • Complex to Build • Decentralized Nature • Interface / Docs Required • Operational Complexity • Transaction Management • Visibility is Difficult
- 6. Microservices at Scale (Excuse the pun) https://www.slideshare.net/aspyker/reinvent-2016-container-scheduling-execution-and-aws-integration
- 9. Who is Talking to Who?
- 10. One Bad Apple…
- 13. The Answer is… • Metrics/Instrumentation – Measure properties of a given system – Alarms and Notifications • Tracing – Observe interactions at a request level – Measure work in time
- 15. What are Metrics? • Metrics are a quantifiable set of measurements of a property for a given system, process, or component. – Performance counters – Instrumentation • Observe behavior • React to changes
- 16. Prometheus • Open-source systems monitoring and alerting project • Cloud Native Compute Foundation (CNCF) hosted project • Originally built by SoundCloud • Data model with time series data • https://github.com/prometheus/prometheus
- 18. Architecture
- 19. Types of Metrics • Counter – only increases in value • Gauge – value goes up or down over time • Histogram – samples observations and counts them over buckets • Summary – histogram plus a summation of value
- 20. Alerts • Create rules based on observed metrics • Alerts trigger actions to be taken – Email – Slack – Webhooks • Why do we care? – Enables dynamic scale up and down
- 21. Prometheus Language Bindings • 15 official and community supported libraries – Go, Java, Python, Ruby, C++, etc • https://prometheus.io/docs/instrumenting/clientlibs/
- 23. What is Tracing? • Enables observability of a given transaction as it moves through a (distributed) system • Allows visualization of which microservice instances are involved • Tracks the path through the software stack + time metrics
- 24. Jaeger • Open-source distributed tracing system • CNCF hosted project • Originally built by Uber • OpenTracing compatible • Root cause and observe performance • https://github.com/jaegertracing/jaeger
- 26. Architecture
- 27. Traces and Spans
- 28. Jaeger Language Bindings • 5 official and bunch of community supported libraries – Go, Java, Python, node, C++ – http://jaeger.readthedocs.io/en/latest/client_libraries/
- 29. Metrics vs Tracing • Metrics – Gives a singular per node, instance, or component view of the world – Health checks, performance monitoring, etc – Alerts and reaction to change • Tracing – Follows a single transaction, API call, etc through a given system or application – Think what a stack trace provides except tracing is doing it in a distributed fashion
- 30. Demo
- 31. Demo Time!
- 32. Demo Configuration • Kubernetes 1.7 • Prometheus 2.1 • Jaeger 1.0 • How-to: https://github.com/dvonthenen/proposals/tree/master/2018_SCALE16
- 33. David vonThenen {code} – Dell Technologies @dvonthenen http://dvonthenen.com github.com/dvonthenen Thank You