Industry executives and experts share their predictions for 2022. Read them in this 14th annual VMblog.com series exclusive.
Testing at the speed of Cloud Native
By Matthew LeRay, co-founder, Speedscale
Kubernetes is fast becoming the standard infrastructure technology for Cloud Native software development. While far from simple, cloud native provides a common pattern for breaking large applications into smaller, highly scalable and resilient services. Engineers enjoy more autonomy and ownership, which dramatically accelerates innovation.
Unfortunately, engineering freedom also creates massive reliability headaches once applications reach production. Whereas a single service becomes much simpler to develop, the application’s end-to-end behaviors become incomprehensible. This challenge is part of the reason that only 15% of DevOps leaders report their organization deploys even once per day, a recent CloudBolt Software report shows. The industry needs a way to shift detection left to improve quality.
To keep pace with cloud native deployment timeframes, SaaS companies need modern Observability and QA tools that enable developers to cut time and eliminate errors in the Kubernetes and container-delivery pipeline. Therefore, three major software quality trends will accelerate in 2022:
Observability of code, infrastructure, and data becomes easier, ubiquitous
Community support for OpenTelemetry (OT) and eBPF will make it easier to gain visibility into complex systems. OT provides a common framework for tracking transactions as they move across services. Many common libraries and technologies like Service Mesh will generate OT traces automatically. This capability will provide a high-level view of how information moves through an application.
eBPF will allow engineers to safely dive deep into the infrastructure and applications running on Linux systems. In the past, sophisticated proprietary agents were necessary to view the inner workings of the network and application, but no more. With donations of capable eBPF collectors from companies such as Pixie Labs and Splunk, the adoption of community-driven observability tools will accelerate in 2022.
The great skills shortage drives tools purchases
The US Bureau of Labor Statistics predicts there will be 1.2M unfilled engineering jobs by 2026. Whether or not that’s accurate, it’s hardly controversial to say that the skills shortage of 2021 won’t get any better in 2022.
The shortage is driven in part by the tendency of many tech-forward companies to build tools in-house rather than buy them from third-party vendors. Historically, this attitude helped accelerate innovation because the software could be built to specification, without adaptation. This trend will begin to reverse itself out of sheer necessity due to the skills shortage.
As highly scalable web infrastructures begin to standardize around Kubernetes, maturing general-purpose tools will once again be able to satisfy the needs of most organizations. When weighed against the effort and maintenance required to build and maintain bespoke solutions, the advantages of off-the-shelf services will be too large to ignore.
Traffic replay makes cloud native apps testable
Integration testing in cloud native is nearly impossible. The constant pace of change means engineers cannot keep the necessary tests, mocks, and data up-to-date. For those that try, the Continuous Integration pipeline becomes littered with flaky tests that break constantly, causing the pace of deployment to slow down. Ironically, because services are more interdependent than the monoliths that came before**,** the value of these tests is higher than ever – despite the challenges.
The future of testing is using real transactions, and traffic replay is a key enabler, both for the developer’s desktop and the Continuous Integration system. Technologies like Speedscale record real-world traffic to create realistic and always-up-to-date tests and mocks. Engineers are more productive when they aren’t toiling away on mocks and tests that are quickly outdated. In addition, engineers are typically not very good at predicting exactly how real users will behave. Traffic replay offers a solution to both of these issues.
Using local development traffic replay, developers bring the entire realistic environment onto their desktop without having to build scripts or test environments. The Continuous Integration system can validate the performance and functionality of new versions of services using automatically generated, always up-to-date recordings.
The Cloud Native philosophy is rewriting the rules of software development at a breakneck pace. The year 2022 is shaping up to be the year of continued and proactive testing at the speed of cloud native.
##
ABOUT THE AUTHOR
Matthew LeRay has 20 years’ experience in enterprise application quality and testing, improving the performance of enterprise, on-prem, distributed, and cloud native applications across multiple generations of technology. His first love is debugging golang code but he occasionally takes a break to craft hand-carved guitars.
