Virtualization Technology News and Information
4 Ways Next-Generation Distributed Tracing Drives Intelligent Observability

By Ajay Gandhi, VP Product Marketing, Dynatrace

A user interacts with an application. That interaction is tagged with a specific identifier. That ID follows the interaction as it launches requests to other applications, microservices, containers, and infrastructure. This is how distributed tracing is supposed to generate observability across a technology stack.

The problem? The increasing complexity of multicloud environments is overwhelming conventional distributed tracing, which relies on correlating data from siloed metrics, logs, and traces. This siloed approach results in blind spots that leave organizations with less reliable and comprehensive observability, which affects everything from user experience to IT operations.

Here are four major benefits of an automated, next-generation-approach to distributed tracing with code-level analysis, and how this approach drives the scalable, automatic, and intelligent observability needed for managing complex and dynamic multicloud environments.

1.       Tame cloud complexity

Multicloud complexity is fueled by shifts away from monolithic software architectures toward dynamic applications, microservices, and cloud-native infrastructure. Conventional distributed tracing cannot scale to the needs, scope, or speed of these environments. As organizations accelerate digital transformation and software innovation with containers, Kubernetes, serverless architecture, and OpenTelemetry, distributed tracing needs AI-assistance and continuous automation to keep up.

Only with AI can organizations expect to be able to automatically process billions of dependencies across their software stack in real time. Only with AI are organizations able to continuously monitor that full stack for system degradations and performance anomalies, prioritize precise answers about those anomalies based on business impact, and then address them at their root cause. All of this comes together to multiply IT teams' productivity and enable them to shift their focus toward higher-value projects and innovations that drive better business outcomes. That's what AI and continuous automation bring to the table in taming cloud complexity.

2.       Enrich observability by putting data in context and going beyond metrics, logs, and traces

Distributed tracing relies on metrics, logs, and traces. However, conventional approaches generally treat metrics, logs, and traces as data siloes, which misses the context of the relationships and dependencies among them across the full stack, and misses the in-depth data that comes with that context.

What do I mean by context? Code-level data, metadata, user behavior data, and OpenTelemetry data used to optimize services and traceability across multiple cloud environments. If distributed tracing relies solely on metrics, logs, and traces, it misses all this other contextual data, which makes it much harder to track dependencies, identify and remediate root causes, and optimize performance and observability at scale.

Without that context, the answers you rely on to make critical decisions or automate operations are not precise enough. You may not be able to identify the true root-cause of a system degradation, and your understanding of the user experience will always be incomplete and out of date. When distributed tracing captures that context, on the other hand, you get intelligent observability stocked with greater, more precise insights that you can use to automate, investigate, remediate, and innovate.

3.       Fill in gaps, blind spots, and incomplete traces

Open-source tools like OpenTelemetry can provide flexibility and valuable instrumentation that improves full-stack observability for the latest cloud-native architectures, but these tools have limits. For example, while OpenTelemetry provides traceability across cloud boundaries, it cannot provide data on container or infrastructure health metrics, or code-level details for method hotspots and CPU analyses. These limitations create gaps, blind spots, and incomplete traces, which leave holes in observability.

Automated, next-generation distributed tracing with code-level analysis automatically discovers the relationships among apps, services, and infrastructure, and enriches low-resolution data, generating better, more precise insights. In other words, filling in the gaps, rather than opening new ones, and following traces all the way through the container labyrinth from end to end.

4.       Relieve the pressures of mounting manual work on IT teams

Gartner estimates over three-quarters of enterprises are in the process of digital transformation. IDC predicts that by 2022, 90% of new enterprise applications worldwide will be developed as cloud-native and utilize more open-source technologies and dynamic architectures like containers, microservices, and serverless.

To get distributed traces from many cloud-native and open-source technologies requires that you make manual code changes in many of the services they interact with. Manual tracing requires manual inputs, manual instrumentation, and manual oversight - more tasks that teams need to dedicate time and effort toward.

This approach doesn't just require more manual input; it fails to optimize data and minimize overhead, which holds back automatic scalability and, ultimately, requires IT teams to perform even more manual operations. Next-generation distributed tracing, however, maintains real-time context and automates instrumentation updates, which relieves a lot of this pressure on IT teams, so they can put their skills and expertise toward more innovating and less menial work.

Intelligent observability thrives on a next-gen approach to distributed tracing

A next-gen distributed tracing solution automatically detects and analyzes end-to-end transactions across apps, services, and multicloud environments with near-zero overhead. No stitching together siloed sets of telemetry data just to get low-fidelity metrics. It just works - solving cloud complexity, automatically filling in gaps and blind spots, and providing critical context for decision making and automating operations. Next-gen distributed tracing meets the needs of modern cloud-native environments and provides the intelligent observability organizations need to succeed.


About the Author


Ajay is the VP of product marketing at Dynatrace. Prior to joining Dynatrace, Ajay helped launch other groundbreaking software platforms at Salesforce, Informatica, Shape Security and BEA. Based in Silicon Valley, Ajay is passionate about travel, cycling and having fun with his two daughters.

Published Friday, November 20, 2020 7:04 AM by David Marshall
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<November 2020>