Rootconf 2018

On scaling infrastructure and operations

Distributed tracing with Jaeger at scale

Submitted by Vivek Sridhar (@vsridhar) on Wednesday, 7 March 2018

videocam
Preview video

Technical level

Advanced

Section

Full talk

Status

Confirmed & Scheduled

View proposal in schedule

Vote on this proposal

Login to vote

Total votes:  +6

Abstract

Distributed tracing is a technique for monitoring & profiling systems built on microservices architecture. Distributed tracing is quickly becoming a must-have component in the tools that organizations use to monitor their complex, microservice-based architecture. Jaeger is an open source tool and part of CNCF project released and worked by Uber.

Outline

Will be covering following topics:

  • Introduction to Microservices
  • What is Distributed Tracing and Open Tracing?
  • Using Jaeger to monitor microservices-based distributed systems covering:

      a. Distributed context propagation.
      b. Distributed transaction monitoring.
      c.  Root cause analysis.
      d. Service dependency analysis.
      e. Performance / Latency optimization
    
  • Demo Jaeger with an example from Monitoring perspective.

  • Demo of tracing to collect application metrics in Kubernetes with Jaeger.
  • Using Jaegar at scale with Kubernetes in a microservice environment.

Speaker bio

Vivek is a tech enthusiast with over 11 years experience in the Software Industry. He is currently working as a Developer Advocate with DigitalOcean and has been a Technology Advisor to several tech startups. Previously he was Head of DevOps & QA at Blackbuck and was a DevOps Solution Architect at HCL (Australia) in client engagement and pre-sales roles. Vivek started his career with IBM Rational (INDIA Software Labs) and is passionate about working with software developer communities.

Links

Slides

https://docs.google.com/presentation/d/1yEM8GjksF9XdOO39gqTIXZVYC_f66f03vB05lbMLUOo/edit?usp=sharing

Preview video

https://hasgeek.tv/rootconf/cloud-sever-management-delhi/1435-auto-remediation-at-scale-using-watchers-vivek-sridhar

Comments

  • 1
    Ramanan Balakrishnan (@ramananbalakrishnan) 8 months ago

    Jaeger seems like an interesting project and definitely relevant given the rising popularity of k8s.

    Comparisons to Zipkin and integration efforts required to instrument existing code would be useful to know before diving in.

    • 1
      Vivek Sridhar (@vsridhar) Proposer 8 months ago (edited 8 months ago)

      Hi Ramanan,

      Thanks for the suggestion and would include these in the slides.

      Will divide talk into following:

      1. Microservices & problems in managing microservices at scale (transaction monitoring, root cause analysis, service dependency, performance/latency)
      2. New Solution apart from Log analysis & Metrics - OpenTracing (Jaeger) & Zipkin {will include comparisons}
      3. Implementation (new code {init & start/finish spans} & upgrading the existing code with Jaeger {This is a good suggestion and like it})
      4. Demo - How Ops & Dev can work together to solve the latency issue in the app (Demo using my Not-Uber App)

Login with Twitter or Google to leave a comment