The State of Kubernetes fleet management 2026

The 2026 State of Kubernetes Fleet Management report examines how organizations manage Kubernetes across increasingly complex distributed environments. With Kubernetes now being essential rather than experimental, the challenge has shifted from adoption to operations: running clusters consistently at scale across clouds, regions, and constrained environments.

This year's research draws from 785 completed surveys, each taking approximately 20 minutes, conducted online between March 2 and March 23, 2026. Respondents included IT decision-makers and practitioners responsible for Kubernetes infrastructure, platforms, and operations at organizations across 4 regions (U.S., EMEA, APAC, LATAM), sourced through expert networks and online panels.

Key findings

Kubernetes has scaled, but operations have not kept pace

Kubernetes is operating at scale across complex, distributed environments, but operational maturity has not kept pace with that growth. In the survey, 85% of organizations report fleet growth over the past 12 months, and 70% now run Kubernetes across multiple cloud providers. At the same time, 65% operate clusters across multiple geographic regions, and 65% manage disconnected or air-gapped environments.

Yet this rapid expansion has outpaced many organizations’ ability to manage it. In our research, 43% of organizations report that complexity has increased over the past year. And despite widespread deployment, strategy maturity lags behind, with many organizations still developing or refining their fleet management approach.

The result is a widening gap between the scale of Kubernetes environments and the operational discipline needed to manage them.

Visibility and governance are established, but not sufficient

Most organizations have made progress in improving visibility and governance. However, consistency and enforcement remain incomplete. Only 17% of organizations report real-time visibility across all clusters, while another 46% have centralized visibility that still falls short of end-to-end insight.

On governance, 75% of organizations have defined fleet-wide standards, and 70% report consistent policy enforcement. But these numbers mask a deeper issue: 75% also regularly override policies to address operational needs.

In other words, the structures exist on paper, but they don’t hold in practice.

This gap between defined policy and enforced practice leaves organizations exposed, particularly as environments grow more complex and distributed.

Lifecycle management is where many Kubernetes operations falter at scale

Lifecycle management—the day-to-day work of upgrading clusters, applying changes, and maintaining consistency—is where operational strain is most visible. Most organizations have introduced some automation, but manual effort remains pervasive. Only 13% report fully automated cluster upgrades, and fewer than half use fully automated, declarative change processes.

The consequences are significant. Three-quarters (76%) of organizations experience configuration drift at least occasionally, 91% report moderate to high effort to keep clusters aligned, and 4 out of 5 issues are related to configuration or change management.

Lifecycle management is the most critical and most inconsistent area of Kubernetes operations, and the primary force behind operational risk.

These challenges are not just operational annoyances; they carry a direct business cost. We found that 62% of organizations report that Kubernetes complexity is actively slowing or delaying business initiatives, reinforcing that this is now an execution issue, not just an infrastructure concern.

Operating model determines outcomes

Not all organizations struggle equally. The research reveals that differences in operating model maturity lead to materially different outcomes. Organizations were scored across 4 maturity tiers: reactive operators (14%), operationally stretched (31%), structured fleet operators (40%), and policy-driven fleet operators (15%).

The gap is stark. Confidence in managing Kubernetes predictably ranges from just 12% among reactive operators to nearly 90% among policy-driven fleet operators. More mature organizations also report fewer disruptions, greater standardization (82% mostly or fully standardized), and stronger alignment between leadership expectations and day-to-day reality.

Maturity is not theoretical. It indicates measurable differences in control, predictability, and business impact.

Future complexity will amplify existing gaps

Looking ahead, the factors shaping Kubernetes complexity are converging rather than stabilizing. AI adoption, continued cluster growth, multicloud expansion, and evolving regulatory requirements top the list of expected forces over the next 1-2 years. At the same time, the most complex operating environments (edge, remote, and air-gapped deployments) are growing fastest.

Nine out of 10 fleet decision-makers and operators expect AI to have an impact on their Kubernetes operations. To prepare, organizations are investing broadly across governance, lifecycle automation, and visibility.

But readiness is not evenly distributed. Policy-driven operators are more than 6 times more likely to feel prepared for future pressure than reactive operators.

Future complexity will not introduce new challenges as much as it will intensify existing issues, particularly around governance, lifecycle management, and operational consistency.

As environments have grown in scale and diversity, operational maturity has not kept pace. The result is a widening gap between what organizations are asking Kubernetes to do and their ability to manage it consistently.

Kubernetes is everywhere

The research makes it clear that Kubernetes has become central to how organizations operate. Results show that 79% of respondents run infrastructure or platform services on Kubernetes, 72% use it for development or test environments, and 71% are running data, analytics, or AI/machine learning (ML) workloads. Internal business-critical applications (67%) and customer-facing applications (63%) round out the picture.

This is not a tool being used at the margins. Kubernetes is supporting the workloads that matter most, and as that footprint expands, tolerance for inconsistency drops. Among the most mature organizations (policy-driven fleet operators), AI/ML workload adoption rises to 84%, signaling that more advanced use cases correlate with more structured operational approaches.

The breadth of workloads running on Kubernetes today raises the stakes for fleet management. Inconsistency is no longer a nuisance. It is a business risk.

Environments are already complex and continue to grow in complexity

These workloads are not running in simple, contained environments: 70% of organizations use multiple public cloud providers for Kubernetes, 65% operate clusters across multiple geographic regions, and 65% manage clusters in disconnected or air-gapped environments.

In other words, multicloud, multiregion, and constrained environments are the norm, not the exception. Each of these dimensions adds a layer of operational complexity, and most organizations are dealing with all 3 simultaneously.

Multienvironment Kubernetes operations are the standard, not the edge case.

Fleets are rapidly expanding 

On top of this complexity, fleets are getting bigger, with 85% of organizations reporting that the number of Kubernetes clusters they manage has grown over the past 12 months. More than 1 in 4 (26%) report significant increases. Decline is virtually nonexistent at just 1%.

Growth is not created only by organic demand. Most organizations (53%) have integrated Kubernetes environments due to a merger, acquisition, or divestiture in the past 3 years. These external events inject additional complexity. They bring inherited configurations, unfamiliar tooling, and clusters that were built to different standards, all of which need to be absorbed into an existing operating model.

Fleet growth is nearly universal, spurred on by both organic demand and external events. Each new cluster widens the management surface and increases the cost of inconsistency.

Strategy maturity is not keeping pace with environmental complexity

Despite the scale and complexity of their environments, many organizations are still developing the strategic foundation to manage them. Only 26% report having a well-defined and mature fleet management strategy. The largest group (51%) says they have a strategy but it is still evolving. And 22% are further behind: 11% are in the early stages of developing a strategy, 9% are still exploring the need for a strategy, and 2% have no strategy at all.

In short, most organizations are running Kubernetes at scale without a fully mature plan for managing it.

This gap becomes more significant as complexity grows and 43% of organizations report that the complexity of managing Kubernetes has increased over the past 12 months. Notably, this figure drops to 25% among policy-driven fleet operators, suggesting that more mature operating models are better able to absorb complexity without compounding it.

The implication is clear: organizations that have not yet formalized their fleet management approach are likely to feel the weight of growing complexity sooner and more acutely than those that have.

Organizations have not been standing still. Most have made meaningful investments in visibility, governance, and policy infrastructure for their Kubernetes environments. Standards are widely defined, visibility is in place across most clusters, and enforcement is generally reported as consistent.

But the data tells a more complicated story. Visibility is rarely real-time. Policies that are defined on paper are routinely overridden in practice. Tooling remains fragmented. And fewer than half of organizations are confident they could demonstrate compliance if audited today. Progress is real, but it has not yet translated into the consistent, reliable control required by fleet management at scale.

Visibility is established, but gaps remain

Most organizations have achieved some degree of centralized visibility into their Kubernetes environments, but full, real-time insight remains rare. Nearly half (46%) of respondents report centralized visibility across their clusters, and an additional 17% have achieved real-time visibility across all environments. At the other end, 33% have visibility into most clusters but describe it as inconsistent, and 3% report only limited or ad hoc visibility.

Progress is clear, but consistency is not.

While the majority can see into their environments, the quality and timeliness of that visibility varies. Gaps in visibility create operational friction, making it harder to monitor, troubleshoot, and maintain control at scale. This challenge is reflected in the broader data: 59% of organizations report that limited visibility across clusters and workloads is a moderate or major challenge.

Without a unified control plane, you lose the ability to manage at scale, turning what should be a streamlined utility into a complex web of visibility blind spots and repetitive manual maintenance." – Comment from a mid-market organization in Australia managing 2-10 clusters

Core challenges span the full operational surface

Visibility is not the only pressure point. The survey asked organizations to rate a range of fleet management challenges, and the results reveal consistent strain across the board. The top challenges reported as moderate or major are:

• Difficulty meeting security, compliance, or data residency requirements (62%)
• Difficulty enforcing consistent configurations or policies (61%)
• Cluster provisioning that is complex or time-consuming (60%)
• Gaps in Kubernetes or cloud-native skills (59%)
• Limited visibility across clusters and workloads (59%)
• Too many tools or consoles required to manage Kubernetes (56%)
• Manual or error-prone application deployment processes (55%)
• Poor collaboration between operations, SREs (site reliability engineers), and development teams (53%)

What stands out is not any specific challenge, but how evenly distributed the strain is. No single issue dominates. Instead, organizations face a set of challenges, with more than half reporting moderate or major challenges in every category surveyed.

These challenges are not isolated—they are systemic.

Governance standards are widely defined, but enforcement breaks in practice

On the surface, governance appears to be in a strong position with 75% of organizations reporting having fleet-wide standards in place. Of those, 24% report that those standards are enforced automatically, and 51% have them formally defined. Another 23% have standards, but they vary across environments, and just 2% have no defined standards at all.

Enforcement also appears solid at first glance as 70% of organizations report a high level of consistency in policy enforcement across their environments.

But there is a critical contradiction in the data: 75% of organizations also report that they regularly override policies to address operational needs.

Organizations are defining policies and then routinely breaking them.

This is not a failure of intent. It reflects the reality of managing complex, distributed environments where operational pressure regularly outpaces governance structures. When an urgent issue arises or an upgrade needs to proceed, policies become suggestions rather than constraints. The result is an enforcement gap that is difficult to detect from the top but creates real exposure on the ground.

Tooling and environments remain fragmented

Governance challenges are compounded by fragmentation. Only 9% of organizations report using a small number of standardized tools consistently across their Kubernetes environments. The vast majority operate with a mix: 63% use a combination of standardized and team-specific tools, 24% use different tools that vary by team, and 5% rely on custom or DIY tooling.

This fragmentation has a measurable impact on operations. When asked about the negative effects of tool and environment fragmentation, organizations report:

• 70% are affected by differences in configuration across clusters
• 68% by the difficulty of managing clusters across different environments
• 66% by the lack of a single source of truth for cluster state
• 59% by the overhead of switching between multiple management tools

Without a consistent tooling foundation, even well-defined governance becomes difficult to enforce. Fragmentation erodes visibility, introduces inconsistency, and makes every operational task harder than it needs to be.

Compliance confidence is uneven

These issues come to a head around compliance. The survey found that enforcement methods vary widely: 66% enforce policies through periodic reviews or audits, 65% use automatic policy or configuration controls, and 55% rely on manual checks by teams. A troubling 37% report that policies are primarily enforced after issues occur, and 20% say policies exist but are rarely enforced.

The downstream effect is predictable. Only 48% of organizations are very or completely confident they could demonstrate compliance or policy adherence across their Kubernetes environments if audited today.

Among the most mature organizations (policy-driven fleet operators), that figure rises to 84%, reinforcing the link between operational maturity and governance effectiveness.

The gap between defining governance and demonstrating it is significant. For more than half of organizations, an audit today would be an uncomfortable exercise.

The previous chapters established that Kubernetes environments are large, complex, and growing, and that visibility and governance, while progressing, remain inconsistent. This chapter looks at where those gaps have the most direct impact: lifecycle management.

Lifecycle management—the ongoing work of upgrading clusters, applying configuration changes, and keeping environments aligned—is the operational backbone of any Kubernetes fleet. It is also where organizations report the greatest strain. Manual processes persist, configuration drift is widespread, and the effort required to maintain consistency is high. When these processes break down, the consequences extend well beyond the platform team: they slow the business.

Automation is incomplete

Most organizations have introduced some level of automation into their lifecycle processes, but full automation remains rare. When it comes to cluster upgrades, only 13% of organizations report that upgrades are fully automated. Meanwhile, 42% describe upgrades as mostly automated, while 36% have only some automation in place, and 8% still perform upgrades manually.

The pattern is similar for configuration changes across clusters. When changes are needed across multiple environments, 47% of organizations apply them through automated, declarative processes. But 42% still rely on scripts or templates with manual oversight, and 5% apply changes manually to each cluster.

Automation has not eliminated manual effort. Most organizations still rely on partial automation and manual intervention for core lifecycle tasks.

"In the cloud, upgrades are automated and easy but lifecycle management for our bare metal nodes in remote workshops is a huge time sink. We spend hours manually troubleshooting connectivity issues or fixing configuration drift rather than focusing on platform improvements." – Comment from an enterprise organization in Spain managing 11-25 clusters

GitOps-based workflows, often cited as a path toward more consistent change management, are gaining traction but are not yet the norm. Only 13% of organizations have adopted GitOps as their standard approach, while 34% use it broadly across most of their fleet, 32% use it for some clusters, and 15% are still in limited or pilot stages. A small minority, at 6%, do not use GitOps at all.

Drift, effort, and unpredictability define the execution gap

The consequences of incomplete automation show up clearly in execution outcomes. Most (76%) of organizations experience configuration drift at least occasionally, meaning clusters gradually diverge from their intended state through manual changes, staggered upgrades, or inconsistent processes. Among policy-driven fleet operators, that figure drops to 48%, reinforcing that more mature approaches significantly reduce drift.

The effort required to keep environments aligned is substantial and 91% of organizations report that the effort to maintain consistency across clusters is moderate to high. Despite that effort, outcomes remain uncertain: only 60% report that the results of changes across their environments are mostly or highly predictable. Among the most mature organizations, predictability rises to 92%.

Drift is common, effort is high, and outcomes are not fully predictable. This is the execution gap that defines lifecycle management today.

"Config drift kills us—clusters diverge from manual tweaks and staggered upgrades, forcing endless audits and firefighting. No single view of the fleet means hours hunting issues across tools/environments. SRE time wasted verifying sameness instead of innovating." – Comment from a mid-market organization in Brazil managing 2-10 clusters

Most issues trace back to configuration and change

Given these dynamics, it is not surprising that configuration and change management are the leading causes of issues across Kubernetes environments. Four out of 5 issues are related to configuration or change management, as reported by 79% of organizations.

When problems do occur, they are discovered through a mix of channels: 70% through centralized monitoring or alerts, 67% through team-specific monitoring tools, and 62% only when applications or services are already impacted. Another 43% find issues during audits or reviews, and 39% through manual checks.

The fact that 62% of organizations discover issues only when services are already affected underscores the reactive nature of many current approaches. By the time an issue is visible, it is already causing disruption.

Operational issues create security and compliance risk

Lifecycle management challenges do not stay contained within the platform team. Configuration, policy, and management issues regularly cascade into security and compliance risk. Over the past 12 months, organizations report that these issues have resulted in:

• Excess permissions (61%)
• Emergency changes triggered by drift (59%)
• Compliance delays or failures (57%)
• Issues found post-production (55%)
• Internal data exposure (41%)
• External data exposure (39%)

These are not hypothetical risks. They are reported outcomes, experienced at least occasionally by a majority of organizations. The link between inconsistent lifecycle management and downstream security exposure is direct.

When lifecycle management is inconsistent, the effects ripple outward: security gaps, compliance failures, and data exposure. The operational challenge becomes a business risk.

The detrimental effect on business

The cumulative effect of these lifecycle challenges is a measurable drag on the business, with 62% of organizations reporting that the complexity of managing Kubernetes is actively slowing or delaying their ability to achieve business goals.

The toll on teams is equally significant. When asked about the negative influence on the people responsible for managing Kubernetes:

• 65% report time spent responding to urgent or unplanned issues
• 59% report context switching between tools and environments
• 58% report difficulty keeping up with change
• 57% report stress or burnout among team members

"Upgrading Kubernetes versions without breaking half the applications that worked fine last week." – Comment from an enterprise organization in the U.S. managing 11-25 clusters 

Lifecycle management is not just an operational bottleneck. It is the primary point where Kubernetes complexity translates into adverse business effects, team strain, and delivery delays.

The previous chapters have shown that Kubernetes environments are large and growing, that visibility and governance remain inconsistent, and that lifecycle management is under strain. But these challenges are not experienced equally. Some organizations navigate them far more effectively than others.

The difference is not the technology. It is the operating model.

This chapter examines how differences in structure, ownership, standardization, and execution discipline lead to materially different outcomes, from confidence and control to disruption frequency and leadership alignment. The data draws a clear line: organizations with more mature operating models achieve better results across every dimension measured.

Operating models vary widely

Kubernetes operating models differ significantly across organizations in ownership, tooling, standardization, platform maturity, and skills capacity. The research measured 6 dimensions:

• 55% report clear ownership of Kubernetes fleet management
• 62% are still at least partially relying on DIY management approaches
• 79% have a defined or established platform engineering function
• 82% describe their environments as mostly or fully standardized
• 57% report that internal skills and capacity are adequate but stretched
• Only 15% operate Kubernetes and virtual machines (VMs) in a unified management model

What stands out is the contrast between surface-level progress and underlying fragmentation. Most organizations have platform teams, most have standards, and most report some level of ownership. But nearly half lack clear ownership, more than 6 in 10 are still partially DIY, and the majority of teams describe their capacity as stretched. The structures are in place, but they are not yet delivering consistent operational control.

Kubernetes operating models look mature on paper. In practice, ownership gaps, DIY tooling, and stretched capacity limit what that maturity can deliver.

Maturity defines structure, standardization, and capability

The differences become sharper when viewed through the lens of operational maturity. Across every dimension of the operating model, organizations at higher maturity levels report stronger foundations.

Platform engineering adoption illustrates the pattern. Among reactive operators, just 42% have a defined platform engineering function. That rises to 72% among operationally stretched organizations, 92% among structured fleet operators, and 97% among policy-driven fleet operators.

Environment standardization follows the same trajectory. Only 34% of reactive operators provide standardized Kubernetes environments or guardrails to application teams, compared to 76% of operationally stretched, 98% of structured, and 98% of policy-driven organizations.

Skills strain tells a nuanced story. Among reactive operators, 58% report that skills and capacity are adequate but stretched. That figure rises to 73% among operationally stretched organizations, then drops to 57% among structured operators, and just 24% among policy-driven operators.

More mature organizations are not just better resourced. They have built operating models that reduce the strain on their teams, not just distribute it.

Confidence and control scale with maturity

The operational payoff of maturity is clear. Confidence in managing Kubernetes predictably at current scale ranges from 12% among reactive operators to 89% among policy-driven fleet operators. That is not a marginal gap. It represents fundamentally different operating realities.

The pattern holds across specific execution capabilities. When comparing reactive operators to policy-driven fleet operators:

• 100% of policy-driven operators can quickly identify and resolve issues, versus 71% of reactive operators (+29 point gap)
• 95% report that operations are under control, versus 66% (+29)
• 95% say security and fleet management are aligned, versus 58% (+37)
• 97% can deploy updates without disruption, versus 67% (+30)
• 99% can predict the impact of changes, versus 62% (+37)

These are not small differences. They describe 2 fundamentally divergent experiences of operating Kubernetes. At 1 end, teams are firefighting, uncertain, and reactive. At the other, operations are predictable, controlled, and repeatable.

Policy-driven fleet operators do not just perform better on every measure. They operate in a different mode entirely, which is defined by control rather than reaction.

Disruptions decrease sharply with maturity

The impact on service reliability is direct. Over the past 12 months, 73% of reactive operators and 77% of operationally stretched organizations report that Kubernetes-related issues have disrupted applications or services. Among structured fleet operators, that drops to 57%. Among policy-driven fleet operators, it falls to 28%.

Reactive operators are nearly 3x more likely to experience service disruptions than policy-driven operators.

"Keeping configs consistent across 100+ clusters is a nightmare. Every little drift causes outages, and manual fixes eat up hours every day." – Comment from an enterprise organization in the U.S. managing 51+ clusters

______

"Upgrading Kubernetes versions without disrupting telco workloads is like playing Russian roulette with patches. Testing on sporadic clusters takes an eternity." – Comment from an enterprise organization in Australia managing 51+ clusters

Leadership alignment reflects operating maturity

Maturity also shapes how well leadership expectations align with the reality of managing Kubernetes. Among policy-driven fleet operators, 87% report that leadership is very or completely aligned with day-to-day operational reality. Among reactive operators, that figure drops to 11%.

The gap is significant. When leadership expectations are misaligned with operational reality, it creates friction: unrealistic timelines, underinvestment in infrastructure, and a disconnect between strategy and execution. When alignment is strong, organizations can plan, prioritize, and invest with clarity.

Operating model maturity does not just improve operations. It closes the gap between what leadership expects and what teams can deliver.

What organizations say success looks like

When asked to describe what successful Kubernetes fleet management would look like for their organization, respondents consistently converged on the same themes: centralized visibility and control, consistency across environments, automation that reduces manual effort, and reliable, stable, secure operations.

"Good Kubernetes fleet management means I can see every cluster at a glance, push updates safely without chaos." –Comment from a mid-market U.S. organization managing 26-50 clusters

______

"Only 1 application to manage all our Kubernetes clusters with the same standards, security, etc. to simplify operations and ensure consistency." – Comment from an enterprise organization in Spain managing 26-50 clusters

______

"Success would be having a unified control plane that lets our small DevOps team manage all 1,000 nodes without manual patching." – Comment from a mid-market U.S. organization managing 11-25 clusters

______

These descriptions are not aspirational visions of some future state. They are practical descriptions of what policy-driven fleet operators are already achieving today. The gap between aspiration and reality is a maturity gap, and the data shows it is closable.

The challenges described in the previous chapters—inconsistent visibility, fragmented governance, strained lifecycle management, and uneven operating models—are not static. They are about to get harder. The forces shaping Kubernetes complexity are converging, and organizations that have not yet built strong operational foundations will face growing pressure.

This chapter looks at what is coming next: the causes of future complexity, the areas expected to bear the greatest strain, where organizations are investing, and the role AI is expected to play.

Complexity is increasing from multiple directions

When asked which factors would most increase the complexity of managing Kubernetes over the next 1-2 years, organizations pointed to compounding pressures rather than a single driver.

The top 4:

• Adoption of AI-enabled applications or platforms (47%)
• Growth in the number of Kubernetes clusters (43%)
• Expansion across multiple cloud or infrastructure environments (43%)
• New regulatory, security, and data sovereignty requirements (39%)

Additional contributing factors include increased expectations from development teams, support for edge or remote deployments, greater geographic distribution, geopolitical and regulatory changes, and mergers, acquisitions, or divestitures.

Complexity is caused by multiple forces that are compounding, making Kubernetes environments more difficult to manage consistently at scale.

The most complex environments are growing fastest

The complexity picture becomes sharper when looking at where organizations expect operating conditions to intensify. Edge and remote Kubernetes deployments show the most dramatic expected growth: just 25% of organizations report that these environments impact their operations today, but 56% expect them to within the next 12 months, a 31 percentage point increase.

Disconnected or air-gapped environments follow a similar pattern, growing from 29% today to 41% in the next 12 months (+12 points). Data sovereignty and regional regulatory requirements remain steady at 43%, reflecting an already significant and persistent constraint.

The environments that are hardest to manage consistently—edge, remote, and air-gapped—are the ones growing fastest. This will put additional strain on operating models that are already stretched.

Future pressure will strain existing gaps

The areas organizations expect to be most strained are not new. They are the same areas where this report has already documented challenges:

• Governance, security, and compliance enforcement (51%)
• Managing upgrades, changes, and configuration drift (44%)
• Visibility into clusters, workloads, and configurations (42%)
• Managing Kubernetes alongside other infrastructure types (38%)
• Operational workload and team sustainability (38%)
• Supporting application reliability and performance (34%)
• Coordinating ownership across teams (24%)

Future complexity will not introduce new challenges as much as it will intensify existing ones.

The top 3 areas of expected strain—governance, lifecycle management, and visibility—map directly to the gaps documented in Chapters 2 and 3. This is not a coincidence. Organizations recognize that the areas where they are already struggling are the areas most likely to buckle under additional pressure.

Readiness to handle that pressure varies dramatically by maturity. When asked how confident they are that their organization can handle future pressures without needing to significantly rearchitect how Kubernetes is managed, the spread is wide: 14% among reactive operators, 26% among operationally stretched, 59% among structured fleet operators, and 86% among policy-driven fleet operators.

Policy-driven operators are more than 6 times more likely to feel prepared for future pressure than reactive operators.

Investment is broad, reflecting systemic need

Organizations are responding by investing across a broad set of operational capabilities. No single investment area dominates, reflecting the systemic nature of the challenge:

• Strengthening governance, security, and compliance automation (56%)
• Improving fleet-wide visibility and observability (52%)
• Automating cluster lifecycle management at scale (51%)
• Expanding GitOps-based deployment and change workflows (49%)
• Simplifying management across Kubernetes, VMs, and containers (44%)
• Building internal platform or fleet management capabilities (40%)
• Supporting edge or remote Kubernetes deployments (35%)

The top 3 investment priorities—governance automation, visibility, and lifecycle management—align directly with the top 3 areas of expected strain. Organizations are directing resources toward the capabilities they expect to need most.

Investment patterns mirror the strain forecast. Organizations are not chasing new capabilities. They are investing to close the gaps they already have before those gaps widen further.

AI is expected to support operations, not replace them

Nine out of 10 fleet decision-makers and operators expect AI to impact their Kubernetes operations over the next 1-2 years. Of those, 32% predict a significant impact, while another 59% expect a moderate impact on efficiency and decision-making.

But the data suggests that organizations view AI as an operational accelerator, not a replacement for operational discipline. When asked which AI-assisted capabilities would be most valuable, the top responses focus on improving execution and decision-making rather than full autonomy:

• Faster root cause analysis across clusters and environments (58%)
• Predictive insights to prevent incidents before they occur (56%)
• Automated remediation of known issues, with human approval (55%)
• Decision support and recommendations for operators (45%)
• Fully autonomous operations with minimal human involvement (42%)

The hierarchy is telling. The most valued capabilities are those that help operators work faster and smarter, not those that remove operators from the loop. Even automated remediation, ranked 3rd, is qualified with "human approval."

Organizations are leaning into complexity, not away from it. AI is expected to improve how Kubernetes is managed at scale, but it will not be a substitute for the operational foundations, governance, lifecycle management, and visibility that this report has shown are still being built.

The findings of this report point to a clear conclusion: Kubernetes adoption alone no longer provides a competitive advantage. Success now depends on an organization's ability to operate with control at scale. Below are 3 recommendations that help close the gap between current operations and where performance levels need to be.

1. Formalize your fleet management strategy and operating model

If your organization lacks a well-defined fleet management strategy (as is the case for 74% of organizations in our research), make it a priority to create one.

The data shows that organizations with mature, policy-driven operating models achieve dramatically better outcomes: nearly 90% confidence in managing Kubernetes predictably, fewer than half the disruption rate of reactive operators, and 6 times greater readiness for future pressure. Strategy maturity is the single strongest predictor of operational success. Define ownership, standardize environments, and establish a platform function that can enforce consistency across the fleet.

2. Automate lifecycle management and close the execution gap

Move from partial to full automation of cluster upgrades, configuration changes, and policy enforcement.

Lifecycle management is where Kubernetes operations break at scale. Today, only 13% of organizations have fully automated upgrades, 76% experience configuration drift, and 4 out of 5 issues trace back to configuration or change management. These are not edge cases. They are the primary source of operational risk, security exposure, and business delay. Invest in declarative, policy-driven lifecycle automation, supported by GitOps workflows, to reduce manual effort, limit drift, and make outcomes predictable.

3. Prepare for converging complexity

Build operational foundations now for the complexity that is already arriving: AI workloads, edge environments, and evolving regulatory requirements.

Future complexity will not introduce new challenges as much as it will intensify existing ones. The top areas of expected strain—governance, lifecycle management, and visibility—are the same areas where organizations report gaps today. Edge and remote deployments are expected to more than double their operational impact within 12 months. AI will reshape how Kubernetes is operated, but it will not be a substitute for the governance and consistency that 9 out of 10 organizations are still working to achieve. Organizations that invest in these foundations now will be positioned to absorb what comes next. Those that wait will find the gap harder to close.

Stop reactive firefighting

Join the 15% who scale with predictability

In the survey, only 15% of organizations have reached policy-driven maturity. These leaders achieve 92% predictable outcomes and face 3 times fewer service disruptions.

Don’t let manual effort and configuration drift stall your roadmap. With Red Hat® Advanced Cluster Management for Kubernetes, you have the tools to move from stretched to stable and become more prepared for the pressures of AI and edge workloads.

Learn more about Red Hat Advanced Cluster Management