Implementing application performance management for improved site reliability

Challenge

Our client’s digital business offers applications which are accessed 24x7 by users across North America. Like many companies, they needed greater visibility into their production systems to appropriately manage workloads, and to identify application performance issues or metrics. In this particular case, seasonal workloads and application issues drove variances in system performance and application stability, which in turn had an impact on the client’s business and reputation. In some cases, issues were reported by the customer before the client had knowledge that an issue existed. Visibility into system performance was imperative to ensue issues were quickly identified and remediated.

Solutions

Leveraging CleanSlate’s expertise with Site Reliability and IT Operations, the following solution was proposed and implemented:
• New Relic Cloud Application Performance Monitor (APM) to help bring visibility to the internal performance metrics of the client’s applications. New Relic APM provided a complete suite of monitoring capabilities including support for application profiling, troubleshooting, and workload analysis.
• The New Relic Infrastructure Management module was also implemented to provide complete visibility of the AWS Cloud Infrastructure. This allowed fine tuning of the Docker Containers deployed within AWS Elastic Container Service (ECS) to ensure applications properly scaled up/down according to user activity.
• CleanSlate implemented New Relic Synthetics to monitor the health of application services from various locations across the country. This validated application health and pinpointed geographic locations that may be experiencing issues.

Results & Additional Info

Improved system stability by identifying and resolving key application performance issues, including challenges with memory allocation errors that were occurring during peak hours. Identified and resolved database performance and connection issues that were causing production outages. Optimized cloud compute costs by having data available to fine tune auto-scaling rules of the container clusters.

Ultimately the solution gave the company visibility to identify issues before the customer and shift the mentality from reactive to proactive troubleshooting.