Flux v2.5 Release: Expanding GitOps Capabilities with CEL Integration and GitHub App Auth - Related to cel, teams, hub, flux, expanding
AI in Agile Product Teams

I have been interested in how artificial intelligence as an emerging technology may shape our work since the advent of ChatGPT; see my various articles on the topic. As you may imagine, when OpenAI’s Deep Research became available to me last week, I had to test-drive it.
I asked it to investigate how AI-driven approaches enable agile product teams to gain deeper customer insights and deliver more innovative solutions. The results were enlightening, and I’m excited to share both my experience with this research approach and the key insights that emerged.
Working With Deep Research: A New Level of Analysis.
My experience with Deep Research was remarkably productive. After providing a detailed research prompt exploring how AI transforms agile product development, I received a comprehensive synthesis that went far beyond what I expected from market research handled by an AI agent.
What impressed me most was how the research agent engaged with my initial request, asking clarifying questions about industry focus, development stages, timeframes, and corporation sizes of interest. This collaborative refinement process ensured the final investigation addressed my specific needs rather than delivering generic information. (Other agents are less inclined to do so; Perplexity or Grok, for example.).
Within just 11 minutes, Deep Research compiled findings from 16 insights into a cohesive narrative featuring three in-depth case studies and a thoughtful cross-case analysis. The analysis didn’t just aggregate information — it extracted meaningful patterns and presented actionable insights in a structured, easily digestible format.
(Download the complete analysis here: AI in Agile Product Teams: Insights from Deep Research and What It Means for Your Practice.).
The investigation examined how diverse organizations leveraged AI within their agile frameworks to transform product discovery and delivery:
Lightful: Agile “AI Squad” Powers Nonprofit Communication.
This London-based tech organization formed a cross-functional “AI Squad” with designers, engineers, and product managers working in daily iterations. Rather than adopting AI for its own sake, they identified specific pain points for their nonprofit clients and experimented with AI solutions in short, rapid cycles.
Their most successful innovation was an “AI Feedback” tool that helps nonprofit consumers improve social media posts by providing suggestions with explanations. The solution educated consumers while augmenting (not replacing) human creativity. The team’s agile approach allowed them to quickly adapt when new AI models became available, swapping in improved technology within Sprints.
PepsiCo: AI Uncovers the “Perfect Cheetos”.
PepsiCo employed generative AI and deep reinforcement learning to experiment with Cheetos’ shape and flavor. First, they built a digital simulation of the production process. Then, they trained an AI system to optimize variables like dough moisture, temperature, and machine settings — running thousands of virtual trials far faster than physical lab tests could allow.
The AI-designed “perfect Cheetos” drove a 15% increase in market penetration by aligning product attributes more closely with consumer preferences. PepsiCo combined human expertise with AI experimentation. Domain experts set clear objectives, while the AI explored the solution space extensively, identifying non-intuitive combinations that human R&D might have overlooked.
Wayfair: Generative AI Enhances Customer Visualization.
Wayfair developed “Decorify,” an AI-powered interior design tool that lets shoppers upload photos of their room and describe a desired style. The generative model produces a photorealistic image of the space filled with Wayfair furniture and decor matching that style, with products linked for purchase.
Within months of launch, the tool had generated over 175,000 room designs for consumers. It addressed a critical customer need: helping me envision what furniture would look like in my space. Wayfair treated this as an MVP: launching early, then improving through iterative updates based on user feedback and usage data.
Across these case studies, Deep Research identified recurring patterns that contributed to successful AI integration within agile frameworks. As the study concluded:
“Common threads in our case studies include a relentless focus on customer needs, iterative development to harness AI’s fast improvements, cross-functional teamwork, and careful attention to ethics and data quality.”.
The six key patterns worth highlighting are:
1. AI as an Insight Engine, Not Just an Efficiency Tool.
In all three cases, AI revealed deeper customer insights that shaped product direction — from identifying content quality needs at Lightful, discovering precise product traits consumers love at PepsiCo and revealing style preferences at Wayfair. Organizations leveraged AI to uncover latent needs and patterns, not just to automate existing processes.
2. Customer-Centric, Problem-First Approach.
Successful teams started with customer problems and needs, then applied AI as appropriate — not vice versa. This discipline prevented wasted effort on “cool” AI ideas that don’t move the needle. The question was always: “How can AI help solve this specific customer problem?” rather than “Where can we use AI?”.
3. Agile Methods Amplify AI’s Impact (and Vice Versa).
The fast pace of AI advancement requires the adaptability that agile practices provide. Teams integrated AI work into their work cadence: using short experiments to test viability, Sprints to build AI-driven attributes incrementally, and frequent reviews to assess outcome quality with stakeholders. This created a powerful feedback loop where Agile’s adaptability enabled quick AI piloting, and AI-generated insights informed subsequent iterations.
4. Cross-Functional Teams and Skills Are Essential.
AI projects intersect with data science, engineering, design, and domain expertise. The most successful implementations involved diverse teams with a shared language around AI. This prevented miscommunication and unrealistic expectations, allowing for smoother collaboration and more effective solutions.
5. Human Oversight, Ethics, and Data Quality.
Teams created processes to verify AI outputs and mitigate errors or bias. This included adding QA steps in the definition of done, A/B testing AI decisions against human ones before full rollout, and proactively addressing ethical considerations. Transparency with consumers and ensuring regulatory compliance were essential.
6. Leadership Buy-In and Culture of Experimentation.
Leadership support provided vision and resources, empowering teams to iterate without fear. Setting realistic expectations — not overhyping AI as magic but as a powerful tool requiring refinement — and communicating progress in terms leadership cares about (customer metrics, ROI, competitive advantage) were crucial.
Becoming Obsolete Is a Choice, Not Inevitable.
What strikes me most about these findings is how they challenge the fear narrative around AI for knowledge workers. Many professionals view AI as a threat rather than a paradigm-shifting technology like the printing press, electricity, or the Internet. Yet, these case studies tell a different story.
In each example, human expertise remained essential. AI enhanced human capabilities rather than replacing them. At Lightful, the AI provided suggestions but kept humans in the creative loop. At PepsiCo, domain experts set objectives and guided the AI’s exploration. At Wayfair, the AI visualization tool helped consumers make more effective decisions but didn’t replace the human shopping experience.
These observations suggest that becoming obsolete in the age of AI is a choice, not an inevitability. The practitioners who thrive will be those who learn to leverage AI as a collaborator — using it to uncover insights from unstructured data, simulate complex scenarios, and enhance their decision-making.
As agile practitioners, we’re uniquely positioned to embrace AI. The agile mindset — focused on adaptation, continuous improvement, and delivering customer value — aligns perfectly with the evolving nature of AI technology. Here are three takeaways for your own practice:
Start small, learn fast. Begin with specific customer pain points where AI might offer value. Run experiments in short iterations, gather feedback, and adapt quickly. Build cross-functional AI literacy. Ensure your team has a shared understanding of AI capabilities and limitations. “Understanding” doesn’t mean everyone should become a data scientist, but everyone should understand enough to collaborate effectively. Keep the human at the center. Design AI implementations that augment human creativity and decision-making rather than attempting to replace it. Most applications keep humans in the loop.
Many agile teams are currently missing opportunities to leverage AI for deeper insights — particularly in transforming qualitative data from user research, retrospectives, and customer feedback into actionable patterns. There’s enormous untapped potential in using AI to extract meaning from the rich but unstructured data we already collect.
The future belongs to agile practitioners who can pair human judgment with AI’s analytical power. We can deliver unprecedented value to our clients and organizations by embracing this partnership rather than fearing it. Deep Research is merely a glimpse into this future.
Have you experimented with AI in your agile practice? What opportunities do you see for AI to enhance rather than replace your team’s capabilities? How might we ensure that AI serves our agile values rather than undermining them? Please drop me a line or comment below.
Interacting with date pickers in web applications can be challenging due to their different implementations. Some date pickers allow direct text input......
Manga, the rich and diverse world of Japanese comics, has captivated readers globally for decades. With the rise of digital platforms, accessing these......
Escalating data breach risks and intensifying regulatory guidelines have put organizations' readiness for privacy protection into the spotlight. Until......
Flux v2.5 Release: Expanding GitOps Capabilities with CEL Integration and GitHub App Auth

Flux, the popular GitOps tool for Kubernetes, has released version [website], bringing a host of powerful new capabilities that significantly enhance its functionality and flexibility. This release represents a substantial evolution in Flux's capabilities, with attributes that the community has been eagerly awaiting. Let's explore what's new and how these changes can improve your GitOps workflows.
Perhaps the most transformative addition in Flux [website] is the integration of Common Expression Language (CEL). This integration has enabled several long-requested elements that give customers more control and flexibility in their deployments.
Flux now supports custom health checks using CEL expressions, allowing you to define precisely how Flux should determine the health of your custom resources. While Flux has always performed health checks for core Kubernetes resources, this new capability lets you teach Flux how to verify the readiness of custom resources that don't follow standard Kubernetes conventions.
For example, when managing a fleet of Kubernetes clusters with Cluster API, you can now define health checks to ensure clusters are ready before deploying addons:
apiVersion: [website] kind: Kustomization metadata: name: prod-clusters namespace: infra spec: interval: 30m retryInterval: 5m prune: true sourceRef: kind: GitRepository name: fleet path: "./production" timeout: 15m wait: true healthCheckExprs: - apiVersion: [website] kind: Cluster failed: "[website], [website] == 'Ready').all(e, [website] == 'False')" current: "[website], [website] == 'Ready').all(e, [website] == 'True')"
In this example, Flux will wait for all Cluster objects to reach the Ready state before proceeding with dependent Kustomizations. The community is also maintaining a health check library with CEL expressions for popular custom resources, making it easy to implement common patterns.
The Receiver API now supports filtering resources based on CEL expressions, giving you finer-grained control over which events trigger reconciliations. This allows for more sophisticated automation workflows and reduces unnecessary reconciliation operations.
Another significant improvement is the addition of GitHub App authentication support. This addresses a common pain point in GitOps workflows: the reliance on personal access tokens or SSH keys that require manual rotation.
With GitHub App authentication, you can configure Flux to interact with GitHub repositories using an identity not tied to a specific user account. This approach provides superior security posture and simpler credential management.
Setting up GitHub App authentication is straightforward using the new Flux CLI command:
flux create secret githubapp github-auth \ --app-id=1 \ --app-installation-id=2 \ --app-private-key=~/[website].
The resulting Kubernetes Secret can be referenced in GitRepository and ImageUpdateAutomation resources, providing a more robust authentication mechanism for your GitOps pipelines.
Flux [website] introduces also, the ability to enrich event metadata by adding annotations to Flux resources. This feature significantly improves the context provided in notifications sent to platforms like Slack or Microsoft Teams. One particularly useful application is including image tags in notifications when Flux's image automation updates container images.
The Flux ecosystem continues to grow with Flux Operator [website], which introduces one of the most requested capabilities: ephemeral environments for testing and validating changes in pull requests.
This powerful capability allows you to automatically deploy application code or configuration changes from GitHub Pull Requests or GitLab Merge Requests to temporary environments. The Flux Operator can create, enhancement, and delete application instances on-demand based on ResourceSet definitions and the state of the PR/MR.
This feature streamlines the development workflow by providing isolated testing environments that accurately reflect how changes will behave in production, without affecting existing environments.
Flux [website] adds several CLI improvements designed to make troubleshooting easier:
flux debug kustomization --show-vars : Inspect the final variable values by merging inline variables with those from ConfigMaps/Secrets.
: Inspect the final variable values by merging inline variables with those from ConfigMaps/Secrets flux debug helmrelease --show-values : Examine the final Helm values by merging inline values with those from ConfigMaps/Secrets.
: Examine the final Helm values by merging inline values with those from ConfigMaps/Secrets flux create secret githubapp : Generate Kubernetes Secrets for GitHub App authentication.
These tools provide greater visibility into Flux's internal state, making it easier to diagnose and fix issues in your GitOps pipelines.
The release includes several other enhancements:
Fine-grained control of garbage collection with .spec.deletionPolicy.
SOPS support for decryption of Kubernetes secrets generated by Kustomize components.
Ability to modification Git commit statuses from events about Kustomizations that consume OCIRepositories.
To take advantage of these new attributes, you'll need to upgrade your Flux installation to [website] For detailed instructions on using these new capabilities, refer to the official Flux documentation, which provides comprehensive guides for implementing each feature.
FluxCD has gained traction as a powerful GitOps tool for Kubernetes, enabling teams to manage cluster state declaratively through Git repositories. However, depending on your needs, several alternatives offer unique capabilities that might more effective suit your infrastructure.
One of the most popular alternatives is Argo CD, which provides a Kubernetes-native continuous delivery solution with an intuitive UI, multi-cluster support, and advanced progressive delivery capabilities. If you’re looking for a GitOps solution designed for large-scale multi-cluster environments, Fleet (from Rancher) is a lightweight yet scalable option. Meanwhile, Jenkins X integrates GitOps principles into a CI/CD workflow, bridging the gap between traditional automation and Kubernetes-native deployments. For teams already using Spinnaker, its support for GitOps workflows makes it a viable option for multi-cloud continuous delivery.
Choosing the right tool depends on your requirements—whether you prioritize UI-driven workflows, scalability, or deep Kubernetes integration. As GitOps continues to evolve, these alternatives to FluxCD are helping organizations simplify and streamline their Kubernetes deployment strategies.
Meta has in recent times introduced data logs as part of their Download Your Information (DYI) tool, enabling clients to access additional data about their pro......
We're a place where coders share, stay up-to-date and grow their careers....
I don’t know if you’ve noticed, but we’re in an “AI or die!” era right now. Every organization seems desperate to figure out how AI fits into their bu......
Azure AI Foundry Labs: A hub For The latest AI research and Experiments at Microsoft

Microsoft has lately launched Azure AI Foundry Labs, a hub designed to bring the latest AI research and experiments to developers, startups, and enterprises. The corporation started this to bridge the gap between groundbreaking research and real-world applications, enabling developers to explore new possibilities.
Azure AI Foundry Labs offers a range of new assets and experimental research projects that empower developers to create prototypes for new technologies. Among the notable projects are Aurora, a large-scale atmospheric model providing high-resolution weather forecasts and air pollution predictions; MatterSim, a deep learning model for atomistic simulations; and TamGen, a generative AI model for drug design using a GPT-like chemical language model for target-aware molecule generation and refinement.
The Labs also feature ExACT, an open-source project that enables agents to learn from past interactions and dynamically improve search efficiency. Magentic-One is a multi-agent system built on the AutoGen framework that solves complex problems by orchestrating multiple agents. Additionally, OmniParser v2, a vision-based module, converts UI into structured elements, enhancing agents' action generation.
The firm writes in an AI and Machine Learning blog post:
Developers can create prototypes using experimental research in Azure AI Foundry Labs, collaborate with researchers and engineering teams by sharing feedback, and help speed up the time to market for some of the most promising technologies.
Lastly, one of the most recent additions to Azure AI Foundry Labs is BioEmu-1, introduced by Microsoft Research. This deep-learning model is designed to predict the range of structural conformations that proteins can adopt. Another one is Magma, a multimodal foundation model designed to understand and act in digital and physical environments.
In a LinkedIn post of Satya Nadella, John C. Hockinson, an AI & Automation Leader, commented:
AI innovation is moving fast, and Azure AI Foundry Labs is putting cutting-edge research directly into the hands of developers. Access to experimental models and breakthroughs means faster iteration, real-world testing, and accelerated AI adoption.
And similar responses on X where Microsoft MVP John Cook tweeted:
Massive move! Giving devs direct access to cutting-edge AI research is how real innovation happens.
Microsoft is not alone in the hyperscale cloud market with AI exploration for developers through Azure AI Foundry Labs. Competing companies like Google and AWS have also launched similar services.
Google, for instance, has Google Labs, which gives developers access to the latest AI innovations, including Project Astra, Project Mariner, and NotebookLM. Project Astra is a research prototype that explores the potential capabilities of a universal AI assistant, while Project Mariner investigates the future of human-agent interaction, specifically through web browsers. On the other hand, AWS offers PartyRock, a mostly free, low-code tool designed for building generative AI applications. This tool elements a new app search function and allows document processing integration into applications.
The Case For Minimal WordPress Setups: A Contrarian View On Theme Frameworks.
The landscape of web and application development is constantly evolving, with new technologies, frameworks, and methodologies emerging at a rapid pace......
There has been a lot of discussion around multimodal AI lately — how these systems can be built, open source options, small-scale alternatives, as w......
Market Impact Analysis
Market Growth Trend
2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 |
---|---|---|---|---|---|---|
7.5% | 9.0% | 9.4% | 10.5% | 11.0% | 11.4% | 11.5% |
Quarterly Growth Rate
Q1 2024 | Q2 2024 | Q3 2024 | Q4 2024 |
---|---|---|---|
10.8% | 11.1% | 11.3% | 11.5% |
Market Segments and Growth Drivers
Segment | Market Share | Growth Rate |
---|---|---|
Enterprise Software | 38% | 10.8% |
Cloud Services | 31% | 17.5% |
Developer Tools | 14% | 9.3% |
Security Software | 12% | 13.2% |
Other Software | 5% | 7.5% |
Technology Maturity Curve
Different technologies within the ecosystem are at varying stages of maturity:
Competitive Landscape Analysis
Company | Market Share |
---|---|
Microsoft | 22.6% |
Oracle | 14.8% |
SAP | 12.5% |
Salesforce | 9.7% |
Adobe | 8.3% |
Future Outlook and Predictions
The Agile Product Teams landscape is evolving rapidly, driven by technological advancements, changing threat vectors, and shifting business requirements. Based on current trends and expert analyses, we can anticipate several significant developments across different time horizons:
Year-by-Year Technology Evolution
Based on current trajectory and expert analyses, we can project the following development timeline:
Technology Maturity Curve
Different technologies within the ecosystem are at varying stages of maturity, influencing adoption timelines and investment priorities:
Innovation Trigger
- Generative AI for specialized domains
- Blockchain for supply chain verification
Peak of Inflated Expectations
- Digital twins for business processes
- Quantum-resistant cryptography
Trough of Disillusionment
- Consumer AR/VR applications
- General-purpose blockchain
Slope of Enlightenment
- AI-driven analytics
- Edge computing
Plateau of Productivity
- Cloud infrastructure
- Mobile applications
Technology Evolution Timeline
- Technology adoption accelerating across industries
- digital transformation initiatives becoming mainstream
- Significant transformation of business processes through advanced technologies
- new digital business models emerging
- Fundamental shifts in how technology integrates with business and society
- emergence of new technology paradigms
Expert Perspectives
Leading experts in the software dev sector provide diverse perspectives on how the landscape will evolve over the coming years:
"Technology transformation will continue to accelerate, creating both challenges and opportunities."
— Industry Expert
"Organizations must balance innovation with practical implementation to achieve meaningful results."
— Technology Analyst
"The most successful adopters will focus on business outcomes rather than technology for its own sake."
— Research Director
Areas of Expert Consensus
- Acceleration of Innovation: The pace of technological evolution will continue to increase
- Practical Integration: Focus will shift from proof-of-concept to operational deployment
- Human-Technology Partnership: Most effective implementations will optimize human-machine collaboration
- Regulatory Influence: Regulatory frameworks will increasingly shape technology development
Short-Term Outlook (1-2 Years)
In the immediate future, organizations will focus on implementing and optimizing currently available technologies to address pressing software dev challenges:
- Technology adoption accelerating across industries
- digital transformation initiatives becoming mainstream
These developments will be characterized by incremental improvements to existing frameworks rather than revolutionary changes, with emphasis on practical deployment and measurable outcomes.
Mid-Term Outlook (3-5 Years)
As technologies mature and organizations adapt, more substantial transformations will emerge in how security is approached and implemented:
- Significant transformation of business processes through advanced technologies
- new digital business models emerging
This period will see significant changes in security architecture and operational models, with increasing automation and integration between previously siloed security functions. Organizations will shift from reactive to proactive security postures.
Long-Term Outlook (5+ Years)
Looking further ahead, more fundamental shifts will reshape how cybersecurity is conceptualized and implemented across digital ecosystems:
- Fundamental shifts in how technology integrates with business and society
- emergence of new technology paradigms
These long-term developments will likely require significant technical breakthroughs, new regulatory frameworks, and evolution in how organizations approach security as a fundamental business function rather than a technical discipline.
Key Risk Factors and Uncertainties
Several critical factors could significantly impact the trajectory of software dev evolution:
Organizations should monitor these factors closely and develop contingency strategies to mitigate potential negative impacts on technology implementation timelines.
Alternative Future Scenarios
The evolution of technology can follow different paths depending on various factors including regulatory developments, investment trends, technological breakthroughs, and market adoption. We analyze three potential scenarios:
Optimistic Scenario
Rapid adoption of advanced technologies with significant business impact
Key Drivers: Supportive regulatory environment, significant research breakthroughs, strong market incentives, and rapid user adoption.
Probability: 25-30%
Base Case Scenario
Measured implementation with incremental improvements
Key Drivers: Balanced regulatory approach, steady technological progress, and selective implementation based on clear ROI.
Probability: 50-60%
Conservative Scenario
Technical and organizational barriers limiting effective adoption
Key Drivers: Restrictive regulations, technical limitations, implementation challenges, and risk-averse organizational cultures.
Probability: 15-20%
Scenario Comparison Matrix
Factor | Optimistic | Base Case | Conservative |
---|---|---|---|
Implementation Timeline | Accelerated | Steady | Delayed |
Market Adoption | Widespread | Selective | Limited |
Technology Evolution | Rapid | Progressive | Incremental |
Regulatory Environment | Supportive | Balanced | Restrictive |
Business Impact | Transformative | Significant | Modest |
Transformational Impact
Technology becoming increasingly embedded in all aspects of business operations. This evolution will necessitate significant changes in organizational structures, talent development, and strategic planning processes.
The convergence of multiple technological trends—including artificial intelligence, quantum computing, and ubiquitous connectivity—will create both unprecedented security challenges and innovative defensive capabilities.
Implementation Challenges
Technical complexity and organizational readiness remain key challenges. Organizations will need to develop comprehensive change management strategies to successfully navigate these transitions.
Regulatory uncertainty, particularly around emerging technologies like AI in security applications, will require flexible security architectures that can adapt to evolving compliance requirements.
Key Innovations to Watch
Artificial intelligence, distributed systems, and automation technologies leading innovation. Organizations should monitor these developments closely to maintain competitive advantages and effective security postures.
Strategic investments in research partnerships, technology pilots, and talent development will position forward-thinking organizations to leverage these innovations early in their development cycle.
Technical Glossary
Key technical terms and definitions to help understand the technologies discussed in this article.
Understanding the following technical concepts is essential for grasping the full implications of the security threats and defensive measures discussed in this article. These definitions provide context for both technical and non-technical readers.