I had the privilege to host a session at Techstrong Predict 2025 last month. The subject was something that has long been on my mind: how reliable, available, and stable we can expect applications to be today as everything about them seemingly gets reinvented – their data structures, hosting models, connectivity – and, especially, lifecycles and production pipelines.
My prediction, which I discussed with my co-panelists, Hope Lynch and Amanda Razani, is that these factors will come to a head this year in the industry, and 2025 will be the year we begin to talk about app “fragility” in earnest.
To summarize, applications are becoming more interconnected, reliant on external services, and infused with AI-generated code. Software teams have already been under pressure to deliver more at a faster pace, and AI and containerization has increased this pressure markedly. Dependencies, open integrations, not-yet-matured AI use, and a speed-to-market culture, all have good reasons to be in place, but they also increase the risks of app dysfunction, failure, degradation, and operational (production) issues.
There are a number of factors that seem to be coalescing now and justifying this prediction that this is the year of the fragile app.
Apps have been fragile for years, but 2025 will be the year this issue becomes noticed and discussed. A convergence of sophisticated cyber threats, steadily maturing user expectations, and the persistent culture of relentless innovation will bring the result – unreliable apps – into a more widespread consciousness. Think about last year’s CrowdStrike incident where a high-profile brand faced operational setbacks.
Another factor at play is AI-driven uncertainties. Organizations are rushing to integrate AI into apps and app development. Completely normal and understandable! But as Lynch pointed out in our session, some companies are given “a basket of money” to incorporate AI without clear direction, leading to confusion and stress among platform engineering teams. The push for AI-driven development, predictive maintenance, and self-healing applications can improve resiliency—but only if done with a level of intention, structure, and human oversight that hasn’t been understood yet.
Properly executed platform engineering can help mitigate fragility by enabling observability, intelligent monitoring, and dependency analysis, improving the developer experience, response times, and overall application stability. As Lynch emphasized, platform engineers should be seen as partners, not separate teams operating in silos, a state we have only started to achieve.
Another thing to note is that DevOps is at a crossroads. Under the pressure of recent change with no letting-up on the gas, DevOps has become somewhat sloppy and is due for a corrective regrounding. The term has been stretched to mean so many things in the fad to “shift left,” and it already started from a pretty ambiguous definition. Some of its core disciplines, such as release controls, integrated telemetry, and cross-functional collaboration, have eroded or never fully taken. Without stronger DevOps practices, the rapid pace of software delivery will introduce more instability than efficiency.
The session also underscored a fundamental tension in IT: Should organizations invest in hardening their systems and platforms? Or should they accept failure as inevitable and focus on lightning-quick remediation? Obviously, this depends on the particular requirements and functionality of each application, but I fear that we have a problem right now with the rapid-recovery side; a slack in engineering with the expectation of failure and a determination to be able to perform through failure.
Why Does This Matter?
Fragile applications pose risks to goal achievement, competitiveness, and consumer protection. Organizations must decide how to approach this challenge in 2025. Is it the moment to proactively invest in strategies to refine DevOps practices, reinforce platform engineering, and reduce app fragility? Or consciously and deliberately adopt a reactive and ready stance, addressing issues as failures emerge?
It is crucial to see this trade-off clearly whether or not my prediction is correct. Managing application fragility ultimately means, as Lynch put it, “prioritizing resilience over perfection.” IT and DevOps practitioners should at least think about a framework that accomplishes this in a way that reflects their organizations’ culture and approach, balancing allocation of resources (e.g., 20% on release quality, 80% on anticipation).
Also Razani shared an important insight which needs highlighting: Communication is key. This is the heart of DevOps – opening up, and accelerating communication. Whether between developers and platform engineers, IT leaders and executive teams, businesses and customers, or agencies and communities, strong collaborative practices will be a differentiating factor in addressing application fragility effectively.
Organizations that act early will be better positioned to accelerate their accomplishments and profits as well as to react and scale rapidly when needed. Addressing fragility now can prevent costly disruptions in the future.
For deeper insights, I encourage you to watch the full session (free registration required). I thank Hope Lynch and Amanda for the sharp, informative discussion, as well as my village of DevOps and application practitioners and experts who helped me bring this prediction to light.
