How BHOOST took troubleshooting from hours to minutes

As a website hosting company, our biggest technology challenge at BHOOST is trying to figure out whether an issue is on the server or software side. Whenever we have a problem, we have to manually comb through logs. The New Relic free tier helped us make headway on this, without having to spend money upfront. New Relic is open source, which also resonated with us as we founded BHOOST three years ago on open-source platforms.

Simple installation and intuitive learning

We started by installing New Relic on one of our virtual servers. For our system administrator, the process of installing New Relic was almost automatic. You take the code and install it into the server—it doesn’t require hours of setup. I also found it easy to learn the different features of New Relic: what I’m not able to figure out intuitively by using the platform, I can quickly learn using documentation. 

After installation, we began using New Relic for tools like logs, dashboards, and application performance monitoring. We track database queries, CPU and run performance. Database queries allow us to zoom in on slow transactions and understand the underlying issue, while the New Relic REST API shows us the average CPU usage per host for a single application.  If something breaks that isn’t traffic-related, we analyze the queries to identify the problem. 

One consistent challenge of ours was monitoring Elasticsearch—a powerful tool because it allows users to conduct GET requests in real time, and it can function as a NoSQL data store. But as your organization grows and needs to process more data, Elasticsearch can have a hard time scaling alongside it and will be more likely to crash. We wanted to understand why that service was crashing more frequently for one of our customers. As mentioned before, as the volume of data you’re working with increases, it becomes more difficult to find the root cause of an issue. With the New Relic free tier, we were able to check on that server and identify the problem. We found the fix and applied it. Troubleshooting this used to take us hours, now we can now consistently identify the root cause of an issue in minutes.

We also monitor Core Web Vitals (CWV)—metrics established by Google to measure a site’s overall user experience—to give us a clear picture of the speed and performance of our customers’ eCommerce sites. They’re particularly important in our industry, a better customer experience has a direct impact on sales and revenue. Monitoring CWV with New Relic is particularly easy because the CWV quickstart creates a dashboard and immediately begins delivering insights.

Preparing for the eCommerce high season

Nearly 90% of our customers are eCommerce companies. Traffic is seasonal, and retail companies can sometimes be caught off-guard by the surges in traffic they see around Black Friday and the holiday season. Ourselves and customers can use New Relic to prepare for traffic surges by:

• Conducting load tests: We put our infrastructure through tests to understand how it will perform when the wave of customers arrives. We set ambitious targets for customer loads and requests per second, then conduct virtual tests to see how the system responds. We do this well in advance so we have time to scale up infrastructure if necessary.
• Focusing on customer experience: Customers choose between multiple gift ideas and retailers during the holiday season. User experience is decisive as they choose what to buy for friends and family. CWV is your friend. In particular, work on improving First Input Delay, which measures interactivity between user input and browser response.
• Getting a head start on holiday promotions: New content doesn’t have to be live before you understand how it will perform. We trial collateral in synthetic environments and take a proactive approach to resolve any issues. The last thing you want to be doing on Black Friday is adjusting or troubleshooting a new promotion. 

Proactive steps with synthetic monitoring

The New Relic free tier enabled us to get a running start on our observability strategy. We plan on taking a more proactive stance by using synthetic monitoring. This will allow us to provide a lot more value in consulting our customers because we’ll be able to test and identify where they’re using their resources inefficiently. Many eCommerce customers don’t understand that speed isn’t just related to the server. New Relic will help us to explain the different factors that impact performance and how upgrading or downgrading different tools will help them achieve the best possible result.