Virtualization Technology News and Information
Datadog 2018 Predictions: The Year of Context

VMblog Predictions 2018

Industry executives and experts share their predictions for 2018.  Read them in this 10th annual series exclusive.

Contributed by Jason Yee, Technical Evangelist, Datadog

The Year of Context


The term "observability" became a buzzword in 2017, perhaps due in part to its ambiguous definition in the world of IT Operations. But observing computer systems is a solved problem. Observability was a problem 10 years ago when Datadog, Prometheus, and so many other tools didn't exist, but today the monitoring market is full of options and most systems can easily produce a deluge of data. The problem in 2017 wasn't observability, but scalability to consume all of that data.

In 2018, the new challenge will be orientation. [John Boyd's OODA loop], a military concept that is often applied to IT operations, espouses agility and speed. The faster one can iterate the loop-observe, orient, decide, and act-the more likely they are to have success. But often overlooked is the importance of orientation. Boyd wrote that orientation shapes the way we observe, decide and act. In terms of IT operations, orientation influences what data we gather, how we interpret it, and ultimately what we do in response to the data.

If 2018's big challenge is orientation, then 2018's main focus will be to provide context to help engineers and developers orient themselves. My three predictions for 2018 as the year of context:

Broadening of DevOps

Although the term DevOps is a portmanteau of developers and operations, in practice, DevOps is the cultural alignment of multiple teams and roles including QA, security, and even business. A large part of context is understanding where and how you and your goals fit within the wider organization. For organizations that have truly adopted DevOps and not just automation tools, gaining more context will mean bringing even more teams into DevOps. And the we don't need to add more abbreviated roles to create a longer term, those roles have always been a part of the DevOps philosophy!

Presently, most organizations have splintered tools such as Salesforce for the sales team, Marketo for the marketing team, and Datadog for the dev and ops teams. In 2018 we'll start to see tools coming closer together, whether loosely integrated or perhaps even some mergers. Bringing these tools closer together will help provide even more of the context required to make informed decisions. Technical teams typically measure success in terms of uptime and system performance, but using business metrics in operations provides a more accurate metric for success. If your business is online, then your technical metrics are your business metrics. They speak to things like volume of sales, efficiency and customer happiness.

Context through abstraction

We've seen the continued abstraction of systems from early configuration management defining individual servers, to tools like Hashicorp's Terraform defining clusters, to very high level abstraction in Kubernetes's service definitions. Along the way we've seen these abstractions handle relationships between applications and nodes. In 2018, we'll continue to see growing adoption of container and orchestration technologies like Docker and Kubernetes. And with it, we will see new abstractions that define the relationships between systems and provide context for what those systems should do and how they should behave.

Robot helpers

"AIOps" will not replace DevOps. Whether the invented term intends to replace developers with AI or make operations engineers obsolete, it's safe to say that robots will not be taking your job in 2018. Nor will they take your job in any foreseeable future if your job deals with complex systems and requires critical thinking (as all jobs in technology do). AIOps will not happen in 2018 and likely will never happen in the fanciful, fully autonomous way that many like to dream about.

Computers are fantastic at doing menial tasks and finding patterns. Just as automation has helped us reclaim time once wasted on tasks like OS installs and configuration, machine learning will help us reclaim time spent orienting ourselves. At Datadog we've already used it to help customers find anomalies, outliers and even forecast potential issues. In 2018, we'll see even more machine learning-provided context to help both technical and non-technical roles make more informed, context-aware decisions. We will even see more organizations use it to automate simple decision making and trigger basic remediation.


About the Author

jason yee 

Jason is a technical evangelist at Datadog, where he works to inspire developers and operations engineers with the power of metrics and monitoring. Previously, he was the community manager for DevOps & Performance at O'Reilly Media and a software engineer at MongoDB. When he's not speaking at conferences or helping organize them, he likes to spend time on planes "travel hacking" and hunting for interesting regional whiskey.

Published Friday, January 05, 2018 8:08 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!
<January 2018>