How OKTA and uShip Scale their Cloud Services with Wavefront Monitoring – Part 1 of 2

September 6, 2018 Stela Udovicic

Last week at VMworld, I had the pleasure to introduce the latest Wavefront enhancements and co-present with two Wavefront customers, Raleigh Schickel of uShip, and Michael Huang of OKTA. I was fascinated by their stories and how both OKTA and uShip rapidly grew their cloud applications with the help of Wavefront’s cloud-native monitoring and analytics platform. For those who were unable to attend VMworld, I’ll share highlights of our session in this two-part blog.

How uShip Scales Their Cloud Services with Wavefront

Raleigh Schickel, manager of software development at uShip, opened up his talk by introducing his company – uShip. Currently, the uShip online shipping marketplace is present in 19 countries, with over 200 employees, 75 of them in the engineering group, and a DevOps team of five. Raleigh talked about how they’re in the process of transforming their monolithic architecture and adopting microservices. The engineering team at uShip have used continuous deployment since 2011, and currently, they’re deploying 5-10 deploys per day. To enable scaling, they needed to decompose their monolith architecture before migrating to AWS. And they needed to monitor the performance of their microservices.

Before Wavefront, the engineering team at uShip tried to use Graphite, an open source metrics tool, for monitoring of their time series data. As uShip acquired more users and started to have more traffic to their service, the DevOps team found that they could not cope with the growing amount of data. It was difficult to get at metrics quickly, and they struggled with creating proper alerts. They decided to move to Wavefront.

Fast forward to today, currently, uShip engineers are monitoring the health of their cloud infrastructure health, their applications, and their business with metrics. Check out the video below to view Raleigh’s on-show-floor chat with Clement Pang, a Wavefront co-founder.

Upon using Wavefront, it was very easy to ingest applications metrics, and it was easy to set up dashboards. With built-in alerting, “you can just select a metric, click add an alert and away you go”, as Raleigh Schickel puts it. Beside having a comprehensive Query Language to make use of its advanced analytics, Wavefront was also pretty at visualizations. All the uShip engineers have access to Wavefront and any developer can add application metrics with a line of code to deploy. Plus, developers can create alerts, which can be forwarded to Slack, email, and VictorOps.

Raleigh also shared some of the alerts that uShip has created in Wavefront and they include alerting on:

  • The amount of overall traffic – high or low
  • The number of servers in the rotation – enough to handle the traffic or not?
  • Failed login attempts
  • Page load time SLAs
  • Scheduled jobs running
  • Payment processing – are payments being processed?
  • Partner traffic – do the partners have sufficient traffic?
  • And last but not least, what is the health of external services that they rely on

One of the various Wavefront dashboards that Raleigh and his team created shows overall application health: it tracks the number of listings per hour, a number of bids per hour, a number of bids accepted. These are core uShip marketplace metrics and they show if there are any deviations from the norm.

It’s similar to their environments’ health dashboards: traffic is seen as divided between web traffic, API traffic, and how many requests are concurrent. Again, anomalies and deviations from the norm are tracked. In addition, Raleigh’s team monitors their servers’ health issues, CPUs, memory usage, network connections.

The DevOps engineers overlay every code deploy as events which makes it easy to correlate anomalies as a result of new code deploys. Raleigh shows all of that in the Wavefront charts, e.g. how the servers behave during deploys, or even whether there is a deploy in progress. On one occasion, he was able to discover a hacker in the middle of the night just by tracking the failed logins metric. Trying to access multiple accounts, the hacker set off a Wavefront alert and wasn’t able to gain access to anything.

In addition, Wavefront dashboards show the outcome of a change in geocoding, enabling developers to validate it: the success of email sending as they shifted from an internal solution to a SaaS, and the change in service charges computing and its impact on SLAs on most hit pages.

On top of it all, uShip is able to track the impact of new code in production. On one occasion, they showed an unexpected, but the positive achievement of a developer who was trying to improve the performance of specific pages, not realizing that it would affect other pages as well. This developer was able to reduce page load times by more than a second on average across all pages.

I also want to share a nice anecdote during the time Raleigh and I worked on our VMworld session. We talked about the Wavefront Kubernetes integration and how Wavefront can be used to monitor a Kubernetes environment. Within 20 min, Raleigh was able to send important Kuberenetes metrics and have it working. This led him to uncover why one of their Jenkins containers was constantly flapping, something that was hidden from his team for a week.

Raleigh’s team is planning to migrate from nearly all of uShip’s infrastructure to AWS within nine months, also planning to expand Wavefront integrations usage and to further to monitor serverless with the newly released Wavefront AWS Lambda SDKs.

In closing, for engineers new and starting with cloud-native analytics for cloud services observability, Raleigh’s advice is summarized below:

  • Instrument everything, as you never know what you would need
  • Monitor and alert on important things, and have alerts around your SLAs
  • Empower your developers and they’ll deliver better products for your organization

In the part two I will share OKTA’s story from their talk. In the meantime, check out the Wavefront free trial.

Stela Udovicic

Get Started with Wavefront Follow @stela_udo Follow @WavefrontHQ

The post How OKTA and uShip Scale their Cloud Services with Wavefront Monitoring – Part 1 of 2 appeared first on Wavefront by VMware.

About the Author

Stela Udovicic

Stela Udovicic (@stela_udo) is a Director of Product Marketing at VMware leading Tanzu Observability by Wavefront PMM team. Before VMware, while at Wavefront, as Sr. Director, Product Marketing, she led Product, Solutions and Partner Marketing. Before Wavefront, Stela led Product Marketing for Splunk's DevOps, IT Ops, storage, and networking solutions. Stela holds an MSc in Electrical Engineering. She has presented at many major conferences, including Splunk.conf, VMworld, DevOps Days, Cisco Live, RSA, Monitorama, PuppetConf, NetApp Insight, etc.

Follow on Twitter Follow on Linkedin More Content by Stela Udovicic
Previous
Monitor Key Metrics for Azure Services and Workloads with Wavefront
Monitor Key Metrics for Azure Services and Workloads with Wavefront

Public cloud adoption is on the rise and if you look at the major players in the public cloud space, Micros...

Next
Going Old-School with Wavefront (or How I FaaS-ed my way to a JSON Endpoint)
Going Old-School with Wavefront (or How I FaaS-ed my way to a JSON Endpoint)

Departure board (split-flap) at Hannover Hauptbahnhof (Hannover main train station), Germany. Wavefront is ...