An agile DevOps approach is an amalgamation of agile sprints and the integrated teamwork of a DevOps model. As Development and Operations teams integrate with agile practices and groups, production and deployment becomes more efficient. Features, updates, and fixes can be delivered weekly, even daily.
This collaborative advancement has established the practices of Continuous Integration (CI) and Continuous Delivery and Deployment (CD). As a result, agile DevOps teams now run a perfectly smooth and flawless CI/CD toolchain.
Except they don't. Why? Because in the agile DevOps framework, there is a vital piece missing; something that previous approaches to application development did well, but has since fallen by the wayside. That is, the post-delivery portion of the toolchain. Without continuous cloud optimization, the CI/CD toolchain still produces massive inefficiencies and overspend.
The Necessity of Cloud Optimization
Cloud optimization is the key to making sure you don't overprovision your app resources, and overspend on your cloud bills. Cloud apps can have a wide variety of functions, and a plethora of moving parts. Depending on your configuration, these parts can run your application for the better or worse. A finely-tuned app is a company's treasure, but an inefficiently tuned one can waste millions of dollars.
With the right tweaks to resources and parameters, overall app performance can improve and costs incurred can be significantly reduced. However, most companies aren't doing this tweaking. Research reveals that 80% of finance and IT leaders report that poor cloud financial management has negatively impacted their businesses. 69% admit to regularly overspending their cloud budget, by at least 25%.
One cause of this is friction between finance departments and application owners. While the CFO and finance teams lobby for saving as much money and as many resources as possible, application owners hate to even consider reducing resources to the applications, afraid that this will cause performance problems and even application failure.
Furthermore, optimization can be a pain to fold into a release cycle. The roadmap gets too crowded by new features and releases, or engineers might not find performance tuning and optimization all that exciting. However: the most likely reason why cloud optimization doesn't happen? Human limitations.
Human Limits in a Virtually Limitless World
Here's a hard pill to swallow: optimization — real, authentic cloud optimization and performance tuning — is too complex for the human brain.
This isn't to rain on the parade of the achievements of human civilization. We humans are capable of great things. But real cloud optimization is far too complicated for humans to perform. In the era of cloud-native microservice architectures, a simple, 5-container application can possess about 255-trillion resource and parameter permutations. This is simply too many data points for a human to try and work with.
Moreover, knowing which permutations to enact requires two distinctive types of knowledge. The first one is infrastructure knowledge, which should cover all stacks: application runtime, cache, compute, database config, job placement, memory, network, storage, thread management, and so on. The second is knowledge of the application workload itself, and its unique features and demands. It's almost impossible to find someone with true depth knowledge of both these realms.
Even if, by some miracle, you find someone with an in-depth familiarity with both types of knowledge, your next problem is the speed of everything. With the constant bombardment of new code, traffic changes, user growth, and new infrastructure options from cloud providers, there's only so much data a human brain can take.
The Solution to Cloud Optimization
Without the right approach and the right tools, true cloud optimization is never achieved. This is why the best thing most companies can do in terms of “performance tuning” is a basic analysis of cloud provider bills.
The solution? Leveraging artificial intelligence (AI) and deep reinforcement learning.
Achieving maximum efficiency for cloud applications requires making judgements and decisions that are too numerous and fast-moving for the human mind – but that are not too numerous for AI.
Deep reinforcement learning, a form of AI, utilizes neural networks based on the connections of the human brain's neurons. Properly trained and developed, these networks can represent hidden data and allow your CO tool to build a knowledge bank of different configurations, in the same way that the brain develops certain behavioral patterns.
An effective cloud optimization tool that leverages these capabilities can aggregate and monitor an entire system, paying close attention to how every shift and tweak in the settings and parameters affects app performance and cost. This processed information is then fed back to the input end of the neural network over and over again, to continuously compound insights.
Compounded insights mean that the network continuously teaches itself to become better at improving the overall efficiency of the application, examining millions of configurations to identify an optimal combination of resource and parameter settings. All the while, as the agile DevOps team continues to improve upon the application, so does the AI-powered cloud optimization tool improve the application's performance and cost utilization.
With each new iteration, the tool's predictions hone in on the optimal solution, and as improvements are constantly found, they are automatically promoted.
Cloud Optimization: The Future of Agile DevOps
With true cloud optimization, agile DevOps teams unlock cost savings, and users enjoy better app performance and user experience. Even though most cloud applications run with more cost than is necessary, such inefficiencies can be eliminated if organizations combine an agile DevOps framework and AI-driven cloud optimization approaches. Cloud apps may be extremely complex, dynamic, and fast-moving, but that does not mean they can't be hyper-efficient, too.
Michael Olson on the AI+ITOPS Podcast: "I really see AIOps as being a core requirement for observability because it ... applies intelligence to your telemetry data and your incident data ... to potentially predict problems before they happen."
Enterprise ITOM and ITSM teams have been welcoming of AIOps, believing that it has the potential to deliver great value to them as their IT environments become more distributed, hybrid and complex. Not so with DevOps teams. It's safe to say they've kept AIOps at arm's length, because they don't think it's relevant nor useful for what they do. Instead, to manage the software code they develop and deploy, they've focused on observability ...
The post-pandemic environment has resulted in a major shift on where SREs will be located, with nearly 50% of SREs believing they will be working remotely post COVID-19, as compared to only 19% prior to the pandemic, according to the 2020 SRE Survey Report from Catchpoint and the DevOps Institute ...
All application traffic travels across the network. While application performance management tools can offer insight into how critical applications are functioning, they do not provide visibility into the broader network environment. In order to optimize application performance, you need a few key capabilities. Let's explore three steps that can help NetOps teams better support the critical applications upon which your business depends ...
In Episode 8, Michael Olson, Director of Product Marketing at New Relic, joins the AI+ITOPS Podcast to discuss how AIOps provides real benefits to IT teams ...