In this post, Andy Denmark, Co-Founder and CTO of Chairish talks to us about how they’re using SignalFx to create a great shopping experience for their customers.

About Chairish

Chairish is an online peer-to-peer marketplace for high-end home furniture and décor. There is a web and mobile application that facilitate the buyer and seller sides of the experience.

Tell us about Chairish and your team.

Andy: Chairish is about two years old now and we are growing at a healthy rate—we have proven the model and demand for this type of service. In addition to the online marketplace, Chairish also provides the services around getting the piece safely to the buyer’s home, including logistics, payment, shipping and warranties. Most of the team here are designers, curators or marketing within home furnishing. My team is responsible for building, running and maintaining the Chairish web and mobile apps.

Tell us a little bit about the nuts and bolts of your application

Andy: Our application is made of a handful of instances on AWS, using EC2, S3, RDS, SQS, SNS, SES, ElastiCache, VPC, and Route 53.

What was changing that made you look for a monitoring tool?

Andy: When we first started, we ran the buy side on Shopify and I focused the app dev on the sell side with the listing flow, API integration to publish to Shopify and the internal tools and operational workflows to complete a transaction and get the product to the customer. We started moving off the Shopify platform as we grew and and I was building more and more components so I now needed a way to manage.

What kind of things are you monitoring?

Andy: We started with standard stuff like disk space, cpu utilization and load, Nginx timing and stats, and Elasticsearch stats. Our dashboard has evolved to also include more application level charts showing request time, memory use per request, etc.

Did you consider other tools? How did you decide on SignalFx?

Andy: One of the things I love about your approach is that it is open. Instead of closed, proprietary agents that are not well understood, you use open source collection agents that we all know well and can write to. We have built a bunch of custom stuff into our app to send metrics to SignalFx. I did look at open source tools that I have used at previous companies but shied away from building my own because there is both a visible and hidden cost. Technically I am capable of doing it but beyond the physical infrastructure costs there are hidden costs of time to deploy and manage this system to make sure it works ongoing.

Have we been useful for your team?

Andy: It’s more than just a bunch of pretty graphs, it’s a platform for me to build on top of. It is very easy for my dev team to share operations data. In my experience the way ops tools work, they create silos leading to contentious relationships between ops and dev especially when things break and the finger pointing begins. For smaller teams like mine, you’re all on the same team. SignalFx allows for easy and immediate transference of visibility and information within the team.

"SignalFx provides an operational metric system for the whole organization."

Andy Denmark
Andy Denmark Co-Founder and CTO

Could you give us an example?

Andy: Sure. Awhile back we had an issue after one of the code pushes where an element on our product pages was leaking memory. It was a slow leak and only happening for admin viewers but was consuming all available RAM. My initial response would have been that I needed more memory and bigger machines. But instead we added metrics and built new charts in SignalFx so we could poke around in the app and discovered the problem. In the end we ended up downsizing our infrastructure and saved money. My team could instantly see the changes in the charts as we were getting to the issue. Those charts are now part of our standard dashboard and we’ve also configured alerts around them so we know right away if something is off.

How has SignalFx helped the Chairish business?

Andy: The usefulness of your product communicates very well. At a company all hands meeting I gave an overview of our architecture, operations infrastructure and how we keep everything up and running. That’s always interesting especially when most of your company is non-technical. But everyone was pretty blown away by the depth of the data. I just pulled up my dashboard and said “this is what’s going on with Chairish right now….how many pages we are serving and how long the requests are taking…”

About the authors

Aneel Lakhani

Aneel is a marketer. Previously he worked on marketing at other startups, served as a Research Director at Gartner, and did stints at big companies like Cisco and IBM doing everything from sales engineering to product management and large scale systems architecture.

Enjoyed this blog post? Sign up for our blog updates