Miniconf on Cloud Server Management (Chennai)

On costs, scaling and securing cloud servers

NorthStar - Monitoring Microservices using Go

Submitted by Aarthi K (@aarthi-k) on Wednesday, 8 November 2017

videocam_off

Technical level

Intermediate

Section

Crisp Talk

Status

Submitted

Vote on this proposal

Login to vote

Total votes:  +1

Abstract

As we move from a huge monolithic architecture to distributed microservices architecture, it gets more difficult to maintain the performance of each microservice and communication between the microservices.

Each microservice is has it’s own API and a set of standards. A change in one can break another. Generally, this is not caught until it goes to production and something breaks and someone complains about it. By then, there has already been a downtime recorded.

NorthStar provides a bird’s eye view from which you can know how your application is performing and helps you catch failures before your customers do. NorthStar is written in Go and uses Swagger files.

This talk will cover how we built a scalable, high performance system using Go and is intended to provide insight to anyone building a webservice on how to effectively monitor it.

Outline

This talk is around how we built a service which can continuously test and monitor your system from the outside.

  1. Conventional Monitoring
    a. Where does it Fail
  2. How does NorthStar work?
  3. Why Go?
  4. Performance - Goroutines & Channels
  5. Fault Tolerance - Contexts
  6. Improvements

Speaker bio

Aarthi is a Product developer working for the Platforms Team at Freshworks and has worked on a highly scalable real time servicing system using Go. She has also developed NorthStar from the ground up with Go.
When she is not coding at work, she is either reading Jeffrey Archer or trying to cover the huge backlog of articles on Pocket.

Links

Comments

Login with Twitter or Google to leave a comment