Get instant Kubernetes observability—no agents required. Meet Pixie Auto-telemetry

How Deeper Observability Helped Viewpoint See Our Way to the Cloud

7 min read

My Viewpoint colleagues Joe Mastroianni, Director of Cloud Operations, and Jesse Darling, Cloud Engineering Manager, also contributed to this cloud migration story.

 

Moving legacy applications to the cloud can be a bit like trying to fit a square peg in a round hole. You know there’s a way, but it’s going to take some work. That’s the situation we had with two of our single-tenant applications being migrated to a public cloud.

As a services company helping the construction industry connect critical business functions like accounting and project management with field operations, we began our own cloud journey migrating the applications to Kubernetes clusters in the cloud. We had limited visibility into these applications to understand how efficiently they were running, their memory footprints, and if there were bugs. That situation prompted a big push to start instrumenting them. However, with the number of hosts pushing 5,000, our existing monitoring tool would be very expensive, and we wanted more end-to-end observability. So we turned to New Relic for help with this cloud migration.

We had prior experience with New Relic, primarily for instrumenting some of our older cloud services that had recently migrated to a cloud native .NET architecture. Almost immediately, New Relic uncovered bugs that had been causing performance problems, which made a huge difference in terms of performance and uptime for our customers. For our single-tenant applications, we saw an opportunity to use New Relic to instrument the host infrastructure along with the applications themselves and get a similar benefit—at much lower cost than our previous monitoring tool. In fact, we reduced our monitoring costs by 57% just moving from our previous solution to New Relic One.

Simplifying a complex cloud migration

This latest cloud migration endeavor with New Relic involved our ERP applications, Vista and Spectrum. They are 30-year-old, single-tenant applications that have been highly customized over the years. We had to instrument multiple clients, the servers and database, as well as third-party services attached to them.

It was a complex undertaking that was greatly simplified by New Relic One. Having the agents under the hood to go out and grab telemetry data required minimal effort on our part. New Relic provided visibility into our application stacks to know how they’re going to perform as we moved them into the new cloud environment. This visibility was immensely beneficial because otherwise, we’d be flying blind. It would have taken much longer to successfully migrate. In the end, we were so confident with how things were going, we switched everything over to the cloud in the middle of the day without interrupting customers.

Another important benefit of having deeper observability in the cloud is the ability to collect benchmark data on things like QA and performance. That’s how we continually improve our products in DevOps. We’re able to run test environments, tear them down, bring them back up with different types of VMs or node sizes, and measure again, all by instrumenting with New Relic One.

Improving engineering efficiency

Having New Relic One has been especially helpful on our single-tenant applications now running in the cloud. Standardizing on a unified platform to observe both application and infrastructure data makes it easy for an ops engineer or support engineer to know at a glance what server, virtual machine, database, or other component is having an issue, which allows us to respond and remediate the issue faster and more efficiently.

A big factor in helping our engineers be more efficient is the reduction in false positives we’ve seen with New Relic. With our previous tool, we were getting about 3,500 alerts per week. All those alerts made it difficult for the engineers to determine which issues required attention and which were just noise. Since deploying New Relic One, we’ve been getting around 600 alerts per week, the majority of which warrant action. There is real value in being able to identify the true issues that will affect our customers and be proactive in addressing them. We expect to see a measurable improvement in uptime directly attributable to being more proactive on alerts.

A better digital experience for customers

Our goal is always to be as proactive with our customers as possible. We don’t want them to call us because they are experiencing slowness in our applications; we want to spot the warning signs early and remediate the issue before they notice any problems.

One example is a back-end integration service that had performance issues. Based on the data from New Relic One, we were able to pinpoint the issue in some middleware. By pulling out that middleware and moving it onto its own cluster, we figured performance would improve five or six times. Performance actually shot up 10 times. In this instance, customers weren’t calling us about slow performance. We proactively identified the issue and addressed it. Instead of having to call us about a problem, customers actually got better-performing service without asking for it.

We also want to be transparent with customers about the issues we’re working on for them. With New Relic, we’ve been able to tailor dashboards and share them with customers so they can see what we’re working on and if it ties to anything they’ve been experiencing.

Having a committed partner is invaluable

As we continue our journey in helping our clients migrate to the cloud, having a solution like New Relic One will play a key role in helping us shift our own classic applications into being cloud ready. Our customers can’t afford downtime with the ERP and construction management applications they depend on from Viewpoint. It’s what they need to keep projects on schedule and pay employees. As the application provider with SLA commitments, we can’t afford downtime either. That makes having our infrastructure and applications instrumented from end to end even more critical in the cloud.

Just as important as the technology from New Relic is the partnership we have. Everything from providing visibility into the cost structure to holding quarterly reviews and periodic support has been consistently positive and helpful. If we’re ever looking into some new or different ways to use the platform or certain capabilities, we can reach out to our New Relic team and count on getting a fast, thorough response. That’s extremely important to us as we move forward on our cloud journey. It can be a complex and challenging process, so having a committed partner like New Relic is invaluable.

Want to experience the kind of full-stack observability that Viewpoint did, simplifying your cloud migration process? Sign up for a forever free account and begin ingesting up to 100GB per month of data at no cost.