There is evidence to suggest that the ‘next big thing’ coming down the road at CIOs is a significant shift towards cloud-delivered, containerised, serverless micro-services operating across highly agile multi-cloud architectures.
At its recent Perform 2020 conference in Las Vegas, for example, Dynatrace quoted a CIO research report which estimated that 70% of organisations are using micro-services, while 88% are expecting to have them deployed by around the end of this year.
Inevitanly there are others who cast doubt on such suggestions and the associated numbers - a report on two such surveys will be following shortly - which goes to show that the reality is likely to be a mix. Some organizations will be keen to go as far and as fast as possible down those roads, while others will take the exact opposite stance. The majority, however, are likely to fall somewhere in the middle, running pilot projects on elements of their production environment or committing to moving sections of their business to the new environments.
That likely reality may, in practice, prove to be even more in keeping with the argument Dynatrace was using at the conference to justify its latest development efforts. This concerns the subject of management of control of operational environments that are inevitably set to become even more complex. To put it bluntly, they are set to move well beyond the capabilities of mere mortals to manage without the addition of increasingly intelligent assistance.
There is an argument which suggests that those businesses going to one extreme of the spectrum or the other will face the more straight forward management problems. The likely reality of a mixed environment, where users move applications and services, or see a business advantage is dynamically switching them around, makes the management issue far more complex.
It is the company’s contention that this growing problem has not yet been met by the appearance of suitable management tools. This has led businesses to start down a DIY path, a track more often snared with pitfalls and failures than with success, with the most common issue being the creation of solutions that may work now but will not scale for the future.
The Dynatrace pitch at providing a solution for this problem is the latest version of its Software Intelligence Patform, where the goal is to provide extended out-of-the-box observation and monitoring for any cloud-native environment. The platform can now onboard and work with data from services running major public cloud service providers like AWS and Azure, Kubernetes-native events, Prometheus OpenMetrics and Spring Micrometer metrics.
According to Steve Tack, Senior VP of Product Management, the Dynatrace platform can now create custom metrics and events based on a wide range of log data so that observability can be extended to any application, script or process that writes to a log file. This then works with Davis, the company’s AI engine, which provides thresholds and baselining algorithms for all infrastructure performance and reliability metrics that enable extended root-cause analysis capabilities. It also enables organizations to scale infrastructure monitoring to match dynamic changing cloud environments, without resorting to manual re-configuration.
One of the key components Dynatrace sees for this coming infrastructure is services orchestration using Kubernetes, which increasingly seems set to be the widely accepted preference for the task. But according to Tack, this is likely to be another contributor to both operational complexity, and the consequent problems of trying to manage operations effectively.
The Davis AI engine has been extended so that it can now automatically ingest Kubernetes metrics, which led to Tack claiming it is the only observability solution that can provide precise answers about issues and anomalies across the full platform to individual workloads stack. The data it can work with includes state changes, workload changes and critical events across clusters, containers and runtimes, meaning it can provide a better understanding of all dependencies and relationships across the entire Kubernetes stack.
It can now be used to optimize Kubernetes resource utilization as well as identify and resolve performance issues and improve business outcomes proactively. It does this by automatically discovering, instrumenting and mapping heterogeneous container technologies within Kubernetes environments, including Docker and CRI-O. The company claims this creates tools capable of managing the largest of containerised environments.
Davis has also been enhanced to enable one-click integration with the most popular web analytics solutions, plus adding the ability to work directly in business-related semantics by directly processing business KPIs. Business users can therefore work directly with Dynatrace using familiar terminology, such as revenue trends, customer conversions and churn, customer loyalty status, geolocation and user satisfaction data. This should deliver a more granular level of business answers and insight, help them directly detect the root cause of business-impacting anomalies.
Mobile devices, particularly as client devices for both staff and customers, are certain to be a significant component in the delivery of business services and end-user experience, and the company has drawn a bead on this as an integral part of managing systems and services end-to-end using the Davis AI engine. It has, for example, added the ability to work with data generated by third-party mobile app components.
The goal here is to deliver precise answers about the health, performance and usage of native mobile apps in real time, making it possible to troubleshoot and resolve issues before user experience and adoption are negatively affected. This is based on multi-dimensional analytics which now include crash reporting, workflow, and granular segmentation capabilities spanning health, performance and usage metrics. Coverage has also been extended to the most popular mobile frameworks and platforms including React Native and tvOS. These augment existing auto-instrumentation for Android, iOS, Xamarin, Cordova and Ionic.
This may, at first glance, appear to be at the ‘spanners and screwdrivers’ end of the IT world, but as the complexity of cloud systems grows, then one simple truth emerges as an important sub-text to the gee-whizz glamour of what the technology can achieve: managing such systems, reliably, in a modern production environment is now just about beyond the wit of any individual human. And while small, well-trained teams can still be effective, any event that steps outside of their experience promptly stalls the management process as the inevitable committee decides what to do.
Automating the management of critical, complex systems is now of primary importance, and the more AI that can be effectively exploited in that purpose the better it is likely to me. So the direction in which Dynatrace is heading here has to be one that most businesses are likely to be obliged to follow.