AppOmni is Minimizing Operational Risk with Elastic Observability's Universal Profiling

AppOmni, a leading cloud security and management platform, offers robust security solutions for businesses dealing with the complexities of cloud-based SaaS applications. To ensure comprehensive and continuous protection for their customers, AppOmni has integrated Elastic Observability's cutting-edge Universal Profiling capabilities. With a vast customer base relying on their services, AppOmni's threat team plays a crucial role in analyzing security threats and operational risks. They work tirelessly to maintain the service's optimal performance and mitigate potential issues. By leveraging Elastic Observability's Universal Profiling, AppOmni's threat team achieves the following:

• Gaining end-to-end visibility across applications, microservices, and infrastructure to identify existing and potential performance bottlenecks, ensuring a seamless customer experience and minimizing service disruptions that could impact revenues.
• Collaborating with developers to not only fix code but also optimize it, resulting in enhanced customer experiences and accelerated product delivery.
• Improving overall team efficiency by streamlining processes and scaling services effectively.
• Reducing cloud costs and CO2 footprint by identifying and optimizing inefficient cloud compute resources, potentially leading to cloud cost savings and reduction in carbon emission.

At AppOmni, Elastic Universal Profiling™ runs continuously in the production environment. It's a crucial tool for code optimization and release validation––creating an ongoing cycle of improvements to ensure peak performance and security for their customers.

In AppOmni's relentless pursuit of ensuring top-notch operational excellence, the threat detection team plays a vital role. They build and operate AppOmni's automated threat detection pipeline that processes billions of customer events per day. In addition to constantly adding new detection capabilities, the team constantly monitors the pipeline for bottlenecks and latency issues that may cause slowdowns in customer event processing. Performance of threat detection pipeline is paramount to the quality of the pipeline.

Their expertise extends beyond just spotting threats operating in and across SaaS applications; they go the extra mile to normalize data, and provide context about threats to their customers. Additionally, AppOmni's threat detection team ships SaaS security alerts about the detected threats to customers, and surfaces these alerts inside the AppOmni application alongside their SaaS security posture and compliance controls.

The AppOmni threat detection team uses Elastic Universal Profiling primarily to identify areas for performance enhancements within the codebase, with a focus on application functions, written in Python and Go. These functions are optimized during development, but when deployed to a production environment at scale, opportunities for optimization appear. This includes improving CPU and memory utilization by identifying portions of the application code for optimization, and evaluating alternative code libraries for performance improvements.

To uphold the highest standards of performance, the threat detection team continuously monitors and reviews the applications in the threat detection pipeline and infrastructure using Elastic Universal Profiling. This proactive approach enables them to expertly manage several key performance indicators (KPIs), including application performance, code optimization, and the prudent reduction of cloud spending.

A notable case of optimization is evident in AppOmni's approach to optimizing web-based microservices used to process events in the threat detection pipeline. Incoming requests are routed to web handlers in the application, each service handling tens of thousands of customer events per second. Ensuring the scalability of the web services while keeping costs in check is a key aspect of AppOmni's optimization strategy.

Using Elastic Universal Profiling, AppOmni achieves the following optimizations:

• Application Code Efficiency: Elastic Universal Profiling analyzes the running web service code, allowing the team to understand how to increase efficiency while reducing execution time to improve overall performance.
• Hardware and Infrastructure efficiency: Elastic Universal Profiling helps identify the most efficient server hardware and infrastructure. This helps the team determine whether to scale up, scale down, or even change the cloud machine type, improving not only overall performance but ensuring the cloud spend is optimal and efficient. In particular, AppOmni uses Google Kubernetes Engine (GKE) in deploying and scaling the AppOmni service. The team was able to reduce unused compute, change out the compute type for more efficiency, and reduce horizontal application scale out.
• Improve Service Performance: Elastic Universal Profiling helps identify optimizations needed for serialization and deserialization, and other processing of pipeline events that might consume significant CPU resources. Universal Profiling helps with identification and comparing performance of alternate solutions. Universal Profiling monitors the service for execution performance and identifies bottlenecks in functions or other areas in the service where it is not performing optimally. The team is able to act on the insights gained from profiling to improve CPU resource consumption and reduce processing time per event. As a result of these optimizations, processing of customer events is faster in AppOmni’s threat detection pipeline.
• Improved Scalability and Responsiveness: Tasks such as customer data synchronization, data analysis, followed by reporting to a customer are scheduled for each customer and run on a periodic basis by AppOmni. Each task is added to an asynchronous task queue based on distributed message passing. Elastic Universal Profiling identified tasks that were consuming more time and resources than expected. This allowed tasks to be optimized, and decomposed into more distributed sets of processing tasks.

AppOmni is using Elastic Universal Profiling to continuously find potential issues and optimization opportunities in the production environment. This has allowed for continual performance improvement with little to no impact on production.

The integration of Elastic Universal Profilin into AppOmni's processes has improved AppOmni's operations by driving optimization efforts and enhancing overall performance. With billions of customer events processed daily, service performance is critical, and the threat team plays a pivotal role in identifying potential improvements. Through Elastic Universal Profiling, the team gains end-to-end visibility, empowering them to streamline code efficiency, optimize hardware and infrastructure, and fine-tune services like web handlers and pub/sub functions. This proactive approach allows AppOmni to deliver faster applications, reduce CO2 emission, and achieve cost-efficient scalability. By continuously leveraging Elastic Universal Profiling, AppOmni stays ahead of potential issues and optimizes its production environment, ensuring a seamless and exceptional experience for its vast customer base.

Don't let performance bottlenecks slow you down. Get started with Elastic Universal Profiling today and unlock deep, whole-system visibility into your code's behavior without instrumentation, code changes, or service restarts. Just deploy the agent and begin optimizing.