Table of Contents
Litmus is an all-in-one email marketing solution that helps customers optimize and personalize every email to maximize ROI and create exceptional brand experiences for every subscriber.
One of Litmus' most popular products, Email Previews, allows customers to test their emails in over 100 real email clients and devices ensuring each email sent is on-brand and error-free.
The technical problem
"Running our own Carbon Graphite interface…wasn’t going to be sustainable.”
Litmus' Email Previews product used to take 2 to 5 minutes to complete a single email test. To be competitive this process needed to be faster. The Litmus team needed to understand which commands were taking the longest time to execute. The team implemented a self-hosted monitoring solution using Graphite.
Before instrumenting Graphite, the team would make changes in their code based on what they thought the issue was noticing a process was slow. This caused problems with performance because, as Brendan Caffrey, VP of Engineering at Litmus, said, “Your gut is not always right.” With a monitoring solution in place, they were now able to quickly track, spot, and solve issues before they become catastrophes.
However as Litmus grew, it became more evident that hosting their monitoring was going to be too costly and time-consuming. It was a struggle to set up self-hosting and they knew that this would be difficult and costly to maintain as they scaled. Brian Feucht, Director of Engineering at Litmus, was typically spending a quarter of his time every week managing their self-hosted observability solution. This was a major problem because of how valuable developer time is that could have been allocated toward other customer-facing or revenue-generating projects.
The solution
"We were running our own Carbon Graphite interface as we built this cluster it became more popular so we realized it wasn’t going to be sustainable… Hosted Graphite could host our metrics cheaper than we could host them in EC2 or AWS ourselves, and we didn’t have to dedicate a developer to that. It was an easy decision to make.”
The Litmus team chose to migrate to Hosted Graphite by MetricFire so that they could focus on making Litmus more profitable as a business and leave running their monitoring to experts. In Brian’s own words, “Hosted Graphite could host our metrics cheaper than we could host them in EC2 or AWS ourselves, and we didn’t have to dedicate a developer to that. It was an easy decision to make.”
Originally, Litmus had custom code sending metrics to their custom sink so sending these to Hosted Graphite instead was easy. They were already using Grafana as well, so they just simply had to import their dashboards into their hosted account.
It’s important for Litmus to have observability and dashboards for their team because they check their dashboards every day to ensure everything is running smoothly.
The result
Litmus has been a loyal MetricFire customer so it is difficult to measure how much time and money Hosted Graphite has allowed them to save. But it is clear that if they wanted to continue self-hosting, this would require a full-time team member dedicated to managing their observability solution.
Litmus now can detect an issue with a hosted email provider before the provider's own status pages are updated. They can also detect other anomalies like if images aren't loading or if a hosted provider has made an unexpected change. This gives them the insight to understand where a problem is coming from and solve it quickly. Most importantly, this means they can increase speed and offer performance as a feature to their customers.
Litmus can send and visualize the data they need by teaming with MetricFire. They have been able to solve real problems, optimise developer workloads and keep their budget under control.
When asked about how our service compares to self-hosting, Brian shared, “[MetricFire is] world’s different; the difference between somebody that specializes in the space and amateurs. It’s very clear that you have a dedicated talented engineering team running your system, and it’s nice to know that you care about metrics as much as we care about email.”
And it’s true. We believe that monitoring should be accessible to all teams, regardless of experience. We work to give our customers meaningful insights into their data and make developers’ lives a little easier.
