Application Impact and IBM Cloud Pak for Watson AIOps

4 min read

IBM Cloud Pak® for Watson AIOps manages your applications, regardless of how they are structured, across multiple management domains, network and storage.

Modern applications and services are built for environments with multifaceted requirements. As the needs of applications grow, complexity increases. This complexity can multiply as, for example, availability requirements are paired with transactional throughput requirements. How can we effectively manage such a complex environment?

The new nexus

Today, applications and services have become the nexus for monitoring and performance, displacing our hardware-centric models. This transformation offers distinct advantages when it comes to monitoring a business.

First, the applications can be aligned with business goals more clearly. For instance, a business application may have a response time goal of less than 150ms or uptime of 99% — these goals flow naturally from understanding how the software relates to the business.

Secondly, the relationships between applications, services and the hardware that powers them are more ephemeral. Container and compute platforms like Kubernetes have traded improved scalability, utilization and resilience for less certainty and granular control over deployments.

Application-centric perspective with IBM Cloud Pak for Watson AIOps

Applications and services are at the manifestation of goals for a business — the contextual center of resources driving execution. But hybrid cloud and cloud-native deployment topologies can turn anomalous log entries, events or metrics into a frenetic game of “connect the dots,” where an ever-evolving container assortment can throw the search for a root cause into disarray.

Using application definitions new in version 3.1, IBM Cloud Pak for Watson AIOps leverages component relationships and machine learning to connect the dots automatically, helping to contextualize event information without the second-guessing. Applications become the hub for analysis of events and information about your applications, services and business.

Application definitions show which resources you want to manage as a collective. Every business, service and application has different expectations about what is included in an application. IBM Cloud Pak for Watson AIOps manages your applications, regardless of how they are structured, across multiple management domains, network and storage. Our application definitions allow you to quickly distinguish between functional components of the application — like network resources and CI/CD — providing greater context and accelerating troubleshooting.

IBM Cloud Pak for Watson AIOps also allows for dynamic application definitions, using criteria that are applied as a filter to the resource topology. An application’s resources can vary over time — based on demand or other criteria — and Cloud Pak for Watson AIOps reflects these changes. This flexibility enables hybrid cloud and cloud-native applications, which often have resources that change over time. When new resources are identified that match the specific criteria, they are automatically added to the application and, likewise, removed when they no longer apply.  The change history of an app is stored over time, letting you view these transitions that are often critical to understanding the context of issues. The result is an accurate view of the application, even in an environment where resources come and go quickly, such as cloud-native applications running on Kubernetes.

The best of both worlds

By offering dynamic grouping, AI-driven application insights now have even more impact. Instead of having to choose between best practices for scaling and accurate models, IBM Cloud Pak for Watson AIOps offers the best of both worlds: horizontally scaling applications with deep observability. Tracing events generated by disparate sources is significantly accelerated with a unified end-to-end view of an application, including a fine-grained record of changes to the application over time.

Compute and container platforms, such as OpenStack and Kubernetes, have transformed application management, but these platforms are not often run exclusively. IBM Cloud Pak for Watson AIOps’ hybrid models allow for a holistic view that allows management of a variety of resource types within the same application, both physical and virtual.

In some cases, specific resources are associated with an application with the intention that it doesn’t change over time. In these cases, application membership can be statically defined to ensure that specific resources are always included. The ability to group arbitrary resources based on dynamic criteria or static membership provides a powerful system for organization where other aspects, such as geographical location or network, may be preferred over application membership.

With both static and dynamic grouping of resources capable of defining applications, a wide variety of applications and services can be represented. A key benefit of this is that teams across an organization can use a common language to discuss the applications that matter. Whether they are maintaining a modern cloud-native application or a service built on hybrid model, the tools within IBM Cloud Pak for Watson AIOps are consistent.

Expanding the model

The value of applications as a hub for monitoring and responding to situations is underpinned by access to key data. The on-going transition to hybrid cloud and cloud-native applications and services has highlighted the need for broad and consistent access to data to power operations. The aspiration to assimilate more data has often been at odds with the continued increase in complexity at production scale, which has made connecting the resources of an application harder and harder. IBM Cloud Pak for Watson AIOps works with a broad cross-section of virtual and physical resources to give consistent access to resources that drive your business.

Consistency in the modelled application has many benefits. A common way to describe and navigate application resources ensures reduced friction as Site Reliability Engineers (SREs) move from application to application during analysis. The ability to manage a diverse set of resources also gives a consistent launching point for delving deeper into additional tools.

IBM Cloud Pak for Watson AIOps brings an equally consistent experience to launching into other Ops, Admin and Management (OAM) tooling. Using these tools in context of the topology of an application gives SREs a more intuitive way to delve further in their analysis.

Conclusion

IBM Cloud Pak for Watson AIOps makes applications the center of gravity for an SRE with simple, customizable application definitions that empower more types of applications and services with AI insights. Adding this power to IBM Cloud Pak for Watson AIOps’ broad catalog of integrations with infrastructure, OAM and data sources, the full breadth of applications are now easier for an organization to manage.

Be the first to hear about news, product updates, and innovation from IBM Cloud