Just a few months ago, observability was a term known only to a few hardcore software engineering types. Suddenly, it's all the rage as application and systems monitoring vendors fall over themselves to launch observability offerings. Now it's the turn of Cisco business unit AppDynamics to expand beyond its application performance monitoring (APM) roots to introduce what it calls a full-stack observability platform.
There are various definitions of observability — cynics suggest that the industry has simply adopted it as the latest in-vogue term for monitoring. To the extent that vendors are rushing to tag their products with the term, there may be some truth in that. But one constant is that the notion of observability extends beyond the application layer to include infrastructure, encompassing cloud platforms and microservices. Another consistent component is that the data is interpreted in real-time and presented in a single pane, where users can drill down into root cause analysis to identify the most important issues and resolve them as quickly as possible. This points to one of the reasons why observability has become a thing — the enterprise IT landscape is becoming simultaneously more complex and yet more critical to everyday business operations, demanding a new generation of performance management tools.
Correlation is key to observability
The AppDynamics announcement must be seen in the context of that changing landscape, says its Chief Product Officer Vipul Shah. Customers are migrating their IT to the cloud, where they need a different toolset than they would have used in a more monolithic architecture. While there are more moving parts to track, many of them are self-healing — for example, if a microservice fails, then Kubernetes will just pop up another instance almost immediately — so you don't need to get alerts about everything. What's important is to focus in on the events that affect the digital experience and its business impact, he explains:
If it is impacting end-user experience and business in a very meaningful way, you want to prioritize that — you want to go and check out what's going on. To us, it's observability as it impacts the business.
This week's announcements are the culmination of a multi-year effort by the AppDynamics team to support cloud-native environments, says Shah, with the ability to correlate multiple data sources one of the most crucial elements: "You cannot achieve observability without correlation."
One of the keys to correlation when dealing with the very large datasets required for observability is a characteristic known as 'high cardinality' — literally, dealing with large volumes of numbers. In the context of a time-series database, this means having to cross-reference multiple indexes each containing a large number of unique values. Pulling all this together and then presenting it as a meaningful visualization is the AppDynamics platform's secret sauce, says Shah:
Finding the needle from the haystack — or high cardinality, choose your favorite set of words — essentially, how can you navigate through a complex set of interactions in applications?
New AppDynamics capabilities
The key points of the announcement are:
- A rearchitecting of the AppDynamics data platform to be able to automatically collect and correlate data from multiple sources alongside its existing agent-based data.
- Ready-to-use frameworks for integrating OpenTelemetry sources such as Jaeger, Prometheus and others.
- New integrations into public and private cloud infrastructure monitoring and telemetry services, starting with Amazon CloudWatch.
- A new visualization interface for monitoring cloud-native applications.
- Integration with Cisco Intersight Workload Optimizer for insight into resource usage and costs.
This combines with existing machine learning algorithms for anomaly detection and root cause analysis, along with the user experience journey mapping which became available earlier this year to provide richer views of web and mobile application performance. Shah also highlighted the impact of integrating ThousandEyes, recently acquired by Cisco, which tracks performance across Internet and cloud connections.
The new capabilities will be most useful for AppDynamics customers either migrating applications from on-premise to cloud, or adopting public cloud for new applications, says Shah. He explains:
A majority of the customers we speak with today are enterprise customers, and these guys, they're all migrating to the public cloud ...
When they migrate their application, it's not always just a lift and shift, it is lift, shift, and continuously transform. That's where, again, AppD comes into play, where as they're transforming, they want to make sure, 'Hey, did we do the right job here or not?'
Observability is software engineering's response to the business need for agility and responsiveness in an always-on digital world. No wonder it's all the rage — it's becoming increasingly crucial to business success in today's economy. That means enterprises need it packaged up so that they can quickly implement it, which creates the market demand that AppDynamics is responding to with this announcement.
To some extent this is a statement of intent rather than a full-fledged platform from the get-go. The basic ingredients are in place but there's a fair bit of work still to do to connect to the various public cloud services and other resources. But there's certainly enough here to give AppDynamics' enterprise customers something to work with as they advance their cloud ambitions.