Application Performance Redefined: Meet the New SignalFx Microservices APM

By Ori Broit

Today, Splunk announced a new milestone release of SignalFx Microservices APM, introducing groundbreaking innovations including: Full Fidelity tracing, AI-Driven Directed Troubleshooting, and open framework instrumentation.

With the Splunk acquisition of SignalFx and Omnition now behind us, we’re excited to announce a new, revolutionary release of SignalFx Microservices APM. By combining the capabilities of SignalFx and Omnition, we delivered the most innovative and advanced distributed tracing solution with open standards based data collection, powering AI-Driven Directed Troubleshooting. These new features go a long way toward our goal of providing best-in-class Observability that enables enterprise DevOps and SRE teams to reduce MTTD and MTTR, improve application uptime, and deliver flawless customer experiences, irrespective of the scale and complexity of their microservices-based applications.

The Application Landscape is Changing

In the past few years, the role of software has become more prominent as it moved from being another part of the business to becoming the face of the business. There is almost no part of modern life that is not conducted without software. At the same time, software development itself has evolved, with new technologies such as cloud, containers, and AI, and also with new methodologies that reduce the time between releases from years to minutes. The changes in software dictate that Application Performance Monitoring (APM) solutions adapt as well:

Observe Everything. The scale and the ephemeral nature of modern environments — coupled with unprecedented speed of innovation and increased importance of software performance — mean that where the next issue might arise is simply unknown, and that any delayed response in addressing that issue will have a negative impact on user experience, brand, and the bottom line. In order to ensure high application performance and availability, ALL the data has to be collected.
AI-driven Analytics. Processing and making sense of all the data generated by modern software is impossible for the human brain. In the modern world, where poor software performance means poor customer experience, old manual processes that are based on trial and error troubleshooting simply will not do. In order to detect issues and find their underlying root cause within seconds, an AI-driven approach is necessary.
Open, Flexible Instrumentation. To meet the needs of the modern speed of innovation, DevOps teams need the most flexible, lightweight tools and programming frameworks, and the creativity of the broad software community. In other words, Open Source frameworks are a must.

Limitations of Existing Approaches

Traditional APM tools are remnants of the past. Aimed at simple, static monoliths, these traditional solutions have not adequately evolved to monitor, explore, and troubleshoot large-scale modern applications composed of microservices, serverless functions, and containerized workloads. The leaders of the old world of APM fall short in every dimension of observability for modern applications. In particular, they’re deficient in three areas that are critical for effectively monitoring, exploring, and troubleshooting:

Partial Information. Traditional APM solutions were built for averages and use head-based, probabilistic sampling. That means they only collect a tiny fraction of the data and, as a result, miss most of the interesting failure and high latency cases that signal the need for troubleshooting. On top of that, traditional solutions were built in silos, which means they can’t effectively connect metrics, events, logs, and traces (MELT). Without the context necessary for today’s distributed applications, operators cannot correlate the performance of their services to underlying infrastructure and business KPIs.
Manual Troubleshooting. Traditional APM solutions were not designed for the scale and complexity of cloud-native applications. As a result, when an issue is discovered, operators need to go through the manual process of analyzing each sampled trace and trying to come up with patterns that seem to make sense. Such manual processes cannot scale in an environment with hundreds of services, thousands of containers, and millions of data points, resulting in mean time to resolution (MTTR) that dramatically increases or fails altogether.
They’re Proprietary. Traditional APM solutions lock in users to heavyweight, proprietary data collection agents that are hard to maintain, cause performance issues, and don’t offer interoperability. They also don’t keep up with the latest programming languages and frameworks, leaving developers with little choice and flexibility in how to build their applications.

Meet The Future of APM and Distributed Tracing

By combining SignalFx and Omnition, we took an entirely different, modern approach to APM, building a solution based on no sampling, and AI-driven analytics, and open standards.

SignalFx Microservices previous generation APM pioneered the use of NoSample™ distributed tracing to observe and analyze every single transaction, perform trace and span metricization, and capture all outliers and anomalies. It was also the first modern APM solution to provide AI-driven analytics and directed troubleshooting to help DevOps teams quickly identify and troubleshoot the offending service or issue. At the same time, Omnition’s distributed tracing solution introduced full-fidelity ingestion with 100% of traces processed and stored in the cloud, with highly granular details (infinite-cardinality exploration) to enable users to perform breakdown analysis on everything of interest.

Both solutions from SignalFx and Omnition already provide the performance and scale to handle large-scale modern applications, rich context that seamlessly links distributed traces to metrics and logs, and full support for flexible, open standards-based data collection, including OpenCensus and OpenTracing – which have merged into OpenTelemetry.

With today’s new release, we’re bringing together the best of these two cutting-edge solutions into a single, APM solution that optimizes around the requirements for modern, cloud-native applications as follows. The net result is a redefined APM solution that is the most flexible, comprehensive, and intelligent on the market.

Use ALL Your Data.
With Full-fidelity NoSample™ Tracing, we’re extending our NoSample ingestion to analyze and store all traces in our cloud. This "Observe Everything" approach combines highly-detailed information and correlates it with the underlying infrastructure, ensuring that you will never miss an error or high-latency transaction.

Make Sense of Your Data, In Seconds.
We’re leveraging advanced AI-Driven Directed Troubleshooting to help you quickly and accurately determine what is contributing to or causing an issue. SignalFx Microservices APM helps you to quickly analyze if a downstream dependent service or underlying infrastructure component is involved in an alert on a service endpoint and helps determine the impact radius of an issue.

Free Your Code.
As founding members and major contributors to OpenTelemetry, we are 100% behind open-standards data collection so our customers can avoid vendor lock-in. Read more about our recent contributions to the open source community here. We also have broad auto-instrumentation support for the most common programming languages and frameworks to get you up and running with minimal effort from your developers.

Taken together, the table below summarizes key features and benefits of the new SignalFx Microservices APM release:

	Features	Benefits
Use ALL Your Data
NoSample Full-fidelity Tracing	Collect, process and store 100% of traces	Never miss an outlier or anomaly Retain full dimensionality of traces Extract full fidelity metrics
Infinite-cardinality exploration	Explore all traces and spans with highly granular details, breaking them down by the container on which they run, version, user, or by any other business logic	Quickly find the root cause of issues Easily understand the impact of any code push
Full-stack correlation with infrastructure metrics	Seamless correlation between infrastructure /integrations & microservices	More quickly troubleshoot service-level issues caused by infrastructure problems
In-context root cause analysis with Splunk logs	In-context access to Splunk logs and events for deeper troubleshooting and root cause analysis	Context-aware workflows across metrics, traces and logs to quickly troubleshoot a performance issue
Make Sense of Your Data, In Seconds
AI-Driven Directed Troubleshooting	Dependency Analysis for dependency-aware incident triage Trace Navigator & waterfall visualization Trace exemplars	Reduce MTTR with real-time, AI-driven analytics and prescriptive approach for troubleshooting Quickly understand the impact radius of an issue and correlate downstream dependencies and underlying infrastructure
Real-time application monitoring & alerting	Auto-populated service/endpoint/performance dashboards Dynamic service maps Low-latency, service-level alerts based on multiple and complex conditions	Instant visualization and interactive visual exploration Fast, easy drill-down into spans Correlated services and infrastructure Shortest MTTD with real-time, accurate, and context-rich alerts with access to representative traces Avoid flappy alerts and alert storms
Trace and span metricization with custom dimensionalization	RED metrics for every unique transaction and every breakdown	Faster, more accurate alerts based on more granular and robust historical baselines
Free Your Code
Open Standards Based Data Collection and Auto-Instrumentation	Auto-Instrumentation for Java, Kotlin, Python, Ruby, Node.js, Go, PHP, .Net Core (Beta) Custom instrumentation Support for open source and open standards: OpenTracing, OpenCensus, Zipkin, and Jaeger; primary contributor to OpenTelemetry Support for Service Mesh (Istio & Envoy)	Rapid time to value with most common languages and frameworks Flexible data collection options to choose from No vendor lock-ins AP

Getting Started with SignalFx Microservices APM

The new release of SignalFx Microservices APM brings Observability to the next level by getting ALL trace data through open standards, and leveraging AI to make sense of it in seconds. This enables DevOps team to innovate and adjust fast to changing market conditions and ensure excellent user experience.

Want to learn more about SignalFx Microservices APM? Visit our website, sign up for our weekly demo, or contact us.

About Splunk

The world’s leading organizations rely on Splunk, a Cisco company, to continuously strengthen digital resilience with our unified security and observability platform, powered by industry-leading AI.

Our customers trust Splunk’s award-winning security and observability solutions to secure and improve the reliability of their complex digital environments, at any scale.