devopsy.kinsta.cloud

Category: Digital Transformation

Common Risks of Outsourcing Software Development, and How to Tackle Them

Both SMBs and large enterprises often choose software development outsourcing over developing software in-house. It is no surprise, as partnering with external developers enables companies to bridge IT talent gaps that cannot be filled internally, avoid time-consuming recruiting and training processes, and eliminate expenses associated with salaries and benefits, eventually accelerating software delivery and reducing development costs.

While software development outsourcing can be highly advantageous from the business perspective, it also introduces various risks, ranging from diminished project control to provider lock-in and security-related issues. Managing these risks proactively is critical to preventing them from escalating, minimizing their impact on the project timeline, budget, and software quality, and ensuring smoother project execution.

In this article, experts from Itransition, an outsourcing partner with over 25 years of experience, outline common risks of outsourcing software development and share recommendations for mitigating them.

Diminished Project Control

Since outsourced teams often operate remotely and without the direct oversight companies typically have over in-house employees, maintaining end-to-end visibility into the development process can be challenging. This lack of visibility can make it more difficult to track project progress and spot bottlenecks in the development process early enough to prevent project delays. To mitigate these risks, a company should establish strong oversight and control mechanisms from the very beginning of the project.

Recommendations:

Ensure your outsourcing partner assigns a dedicated single point of contact (SPOC ) and define strict communication cadences directly in the outsourcing contract to ensure that the provider’s SPOC regularly delivers project status updates.

Additionally, negotiate with a partner about implementing visibility dashboards that automatically pull data from the provider’s toolchain (project management apps, Git repositories, etc.) enabling your stakeholders to monitor project progress and milestones without requesting manual updates.

To maximize transparency and gain evidence-based insights into project progress, require your outsourcing partner to maintain documentation, wireframes, architectural designs, and other project deliverables in a shared, client-accessible workspace.

Security and Privacy Concerns

Software outsourcing partners often require access to clients’ internal systems to perform development-related tasks. For instance, a partner may need to access internal middleware or APIs to integrate software into the client’s existing tech ecosystem. External developers may also require access to CI/CD pipelines to run tests or deploy software updates. However, granting third-party access to corporate systems inevitably introduces various cybersecurity risks, such as data breaches, credential exposure, or intellectual property theft. Taking potential security risks into account and addressing them proactively is essential to avoiding them.

Recommendations:

Add dedicated external-facing roles, such as Partner, Partner Admin, or others, to your existing role-based access control (RBAC) framework to ensure third-party developers can only access systems required to complete their specific tasks.

To prevent permission creep, implement just-in-time (JIT) provisioning mechanisms so that permissions granted to external developers remain active only for a limited period and are automatically revoked when the allotted time expires.

Configure multi-factor authentication (MFA) methods and require outsourcing developers to use it when accessing your company’s internal systems to ensure that developer accounts cannot be exploited by malefactors to breach your IT infrastructure.

Provider Lock-in

Delegating coding, testing, maintenance, and other software development tasks to an outsourcing partner inherently introduces the risk of provider lock-in. This occurs when a company becomes so heavily dependent on a particular partner that switching providers or bringing development processes back in-house becomes too disruptive, technically complex, or prohibitively expensive. To reduce the risk of provider lock-in, consider implementing several useful practices.

Recommendations:

Verify that the contract with a partner clearly defines ownership rights for all intellectual property (software specifications, source code, UI/UX designs, AI algorithms, etc.) to eliminate ownership disputes if you decide to switch to a new software outsourcing provider.

Establish a centralized internal knowledge base containing key project documentation (system diagrams, developer guidelines, etc.) and require ongoing knowledge transfer sessions between in-house teams and outsourced developers. These measures will help you avoid a deficit of internal knowledge and skills in case you decide to transition the development in-house in the future.

It is also highly recommended to negotiate with a partner to provide transition support services for a set period after the contract terminates (several weeks or months) to ensure business continuity and seamless handover to a new agency, in case you decide to switch outsourcing providers.

Final Thoughts

If you plan to delegate some part of your project or entrust its end-to-end execution to an external team, proactively managing the inherent risks of software development outsourcing is essential to ensuring smooth project delivery and its overall success.

The recommendations listed in this article will help you address and mitigate common risks involved in outsourced projects, including lack of project control, security issues, and provider lock-in.

To prevent these and other risks from escalating, it is equally important to choose an experienced outsourcing partner with a proven track record in your particular industry or domain. A reliable partner will help you develop software that meets your project requirements, as well as support knowledge transfer, protect your intellectual property rights, and ensure compliance with relevant data security and privacy regulations.

August 3, 2026
I Learned Traffic Optimization Before I Learned Cloud Computing. It Turns Out the Lessons Were the Same.

I wanted to be a race car driver before I knew what a data center was. I started in traffic, not in the cloud. This was not a childhood dream driven by glamour. It was more practical than that. I grew up in India, and I was always late for school. Not entirely my fault. I have a twin brother, and if one of us was late, both of us were late. So I learned to optimize. I memorized traffic patterns, learned which signals stayed red the longest, and remembered every pothole and shortcut. I did not drive fast for the sake of speed. I drove to arrive exactly when I needed to, with as few surprises as possible. Looking back, this was not about racing. It was about control.

Speed Needs Direction

Speed matters. In racing, it’s the difference between winning and never catching up. But speed without awareness is how you miss a turn. What actually separates good drivers from reckless ones is not how fast they go, but how early they know what’s coming.

Cloud infrastructure has the same problem. Teams are not struggling because they can’t move fast enough. They struggle because systems behave differently in production than they did on paper. Costs rise for reasons no one can fully explain. Performance looks fine until traffic patterns change. Security updates introduce side effects that ripple across environments. Engineers end up reacting to surprises instead of making deliberate choices.

When speed is paired with clarity, it compounds. When it’s not, it creates stress. The difference is whether you understand the system well enough to trust it at full velocity.

Why Cloud Becomes a Black Box

Public cloud platforms are powerful, but they’re not transparent. Pricing calculators assume steady usage while real workloads are anything but steady. Data moves across regions, APIs scale unpredictably, and platform services look convenient until you realize how difficult they are to leave. Over time, teams lose the ability to reason about their own systems. Infrastructure stops being something you design and becomes something you simply tolerate. That’s not an engineering failure. It is a loss of control.

Engineers Do Better When They Have Options

In racing, the moment you lose options, you’re in trouble. The same is true in the cloud. When teams are locked into a single provider or architecture, optimization slows down. Every decision feels risky. Every change feels expensive. Engineers stop improving systems and start working around limitations. Portability restores options. When infrastructure can be cloned, moved, and recreated across environments, teams can compare, test, and choose instead of guessing. This isn’t about multi-cloud for the sake of it. It’s about removing the feeling of fear from decision-making.

Efficiency is About Removing Friction, Not Cutting Costs

Efficiency is often framed as cost reduction, but that misses the point. The real waste in cloud environments isn’t money. It is time spent compensating for systems that can’t adapt. Engineers spend weeks rewriting infrastructure because portability was never part of the original design. When systems behave predictably, teams move faster without rushing. They make better decisions with less stress. They even sleep better, sometimes. I do not, but that’s a separate issue.

You Learn to See Failure Before it Happens

When the recent cloud outages hit, a lot of people treated them as surprises. From the inside, they weren’t. When you spend enough time working with infrastructure, patterns repeat. The warning signs are usually there early, just easy to ignore when things are working well enough.

This reminded me of how I used to ride to school. If you travel the same roads every day, traffic stops feeling random. You notice the small things. A pothole that gets worse after rain. A signal that always backs up at a certain time. A stretch of road where one delay turns into five. You learn where not to accelerate, not because you can’t go faster, but because you know what’s coming next.

Cloud systems behave the same way. Outages rarely come from a single failure. They come from layers of hidden coupling. Services that assume too much about each other. Infrastructure that can’t be moved or tested outside its original environment. When something breaks, it breaks everywhere at once.

The problem isn’t that the cloud is fragile. It is that most teams are flying blind. They can’t simulate failure easily. They can’t move workloads quickly when conditions change. They’re locked into architectures that work until they very suddenly don’t.

Speed matters most in those moments. The ability to clone infrastructure, shift environments, and recover without redesigning everything is the difference between a disruption and a disaster. You can’t avoid every pothole. But you can build systems that let you adjust before you hit it.

The Same Instinct Still Drives Me

I grew up writing code because I wanted things to work better, and I still operate the same way. I can work for days without sleep if the problem is interesting enough. I get restless when systems behave inefficiently. I notice friction quickly, and once I see it, I want to remove it.

Cloud infrastructure shouldn’t feel like traffic you’re reacting to in real time. It should feel like a road you know well enough to adjust before the problem appears.

The future of cloud isn’t going to belong to the loudest platforms or the longest feature lists. It’s going to belong to systems that behave predictably, give engineers real control, and let teams focus on building instead of compensating. That’s the kind of system I am interested in building.

February 27, 2026
Survey Surfaces Lots of Room for Software Engineering Improvement

A survey of 600 software engineers and 600 senior technology leaders across the United States, United Kingdom, Germany, and France finds that nearly three quarters of software engineers say demands on their time make it difficult for them to build new features.

Conducted by Chainguard, a provider of repositories for accessing secure container images, the survey also finds that 93% of engineers said they find writing code and building new features to be rewarding, but they spend only 16% of their week on this type of work. Just over a third (35%) also noted that excessive workload and burnout have become major obstacles to a positive work experience.

Additionally, only a third (33%) of software engineers strongly agreed that they spend the majority of their time on work that energizes them, even though 65% said that the majority of their organization’s common software engineering tasks are either mostly or fully automated. However, 62% also noted that their organization’s tools are not fully integrated into workflows.

Unfortunately, the survey also suggests the first wave of artificial intelligence (AI) technologies is not yet having a major impact on productivity. While 89% of respondents reported AI saves them at least three hours per week, only 28% said they reclaim up to six hours.

Dustin Kirkland, senior vice president of engineering at Chainguard, said the survey makes it clear that the developer experience is suffering as a result of the friction that exists within DevOps workflows. A full 88% of respondents, for example, said switching between tools impacts their productivity, with 44% reporting significant loss of focus due to this context switching. While many low-level tasks have been automated, it’s clear that the level of integration that exists between tools is not all that high, added Kirkland.

Well over a third (38%) of engineers cited too many tedious tasks as a barrier to doing meaningful work, while 38% pointed to the ongoing demands of code maintenance, such as upgrades, patches, and vulnerability management as additional barriers. As a result, software engineers are not experiencing the level of joy that’s needed to keep them from burning out, said Kirkland.

The survey of IT leaders makes it clear that organizations are aware of these issues, with two-thirds (66%) reporting that they worry about retaining engineering talent. Two-thirds of all respondents (66%) said they frequently or very frequently encounter technical debt that affects their ability to deliver work effectively. Nearly half (49%) of all respondents agreed that tool-related issues contributed to a negative developer experience.

It’s not clear what impact AI might have on the developer experience, but as workflows become more automated in the forthcoming agentic era of AI, the overall developer experience should improve. The challenge is that in the short term many of the workflows that DevOps teams use today may be too fragmented to effectively automate, noted Kirkland.

In the meantime, however, it’s clear there is much room for software engineering improvement at a time when expectations about how quickly software can now be built and deployed have already risen sharply.

October 8, 2025
Accelerating Telecom Application Modernization With AI and DevOps

For decades, telecom operators have relied on legacy systems that are highly customized, deeply intertwined and challenging to scale. These systems have supported critical functions such as billing, provisioning and customer management, but in many cases, they are now holding providers back from innovation, agility and long-term growth. As the industry pivots toward cloud-native, data-driven architectures, modernizing these legacy environments has become a business priority.

Today, a combination of generative AI, DevOps practices and modular system design is enabling telecom organizations to transform with greater speed, intelligence and resilience, without destabilizing the core of their operations.

Using GenAI and DevOps to Streamline Legacy Transformation

You’d have to be living under a rock not to recognize that gen AI is proving especially useful in large-scale transformation projects. It can be used to analyze legacy code, generate documentation for poorly documented systems and even propose refactored versions of modules for modern architectures. When paired with DevOps pipelines, this reduces manual effort and helps teams maintain velocity throughout the modernization journey.

For instance, during the migration of legacy code at a major cable provider to the latest Java Springboot standards, GenAI analyzed the existing Ruby on Rails code. It created comprehensive documentation for the migration process to Java Springboot and assisted in the deployment on Anthos P3. The project followed a phased approach: Initially analyzing the legacy code, documenting the requirements and subsequently converting those requirements into architectural artifacts. Windserf (Codeium) was employed to first convert Ruby code into Java and then upgrade it to the Springboot framework. This initiative significantly streamlined onboarding, reduced reliance on individual expertise, increased team velocity up to 30%, improved code quality and mitigated security vulnerabilities through robust DevOps pipelines.

Meanwhile, DevOps provides the structure needed to make modernization iterative and testable. Continuous integration and continuous delivery (CI/CD) pipelines allow teams to test small components in isolation and deploy updates with minimal risk. This is crucial in telecom environments, where even minor outages or performance lags can affect thousands of customers in real-time.

Containerization, infrastructure as code (IaC) and automated testing environments can all play a role in supporting clean deployments as systems are broken down into modular, more maintainable components.

Rationalizing Complex Billing Systems

One of the most complex modernization challenges telecoms face is rationalizing legacy billing systems. These platforms often include decades of custom logic for rating, charging, discounting and revenue recognition. They’re also frequently interwoven with customer care and provisioning tools, making them even harder to decouple.

To move toward commercial-off-the-shelf (COTS) readiness, organizations must first normalize and document existing business logic. AI can assist here by identifying redundant logic paths, surfacing undocumented dependencies and creating test cases that mimic production behavior.

Organizations preparing for COTS should standardize and document their existing business logic. Artificial Intelligence facilitates this by identifying redundant processes, uncovering dependencies and creating test cases. During the transition from old billing systems to a new Telecom Billing COTS platform, business rules are extracted and normalized. The Billing and Product Catalog can integrate with CRM, provisioning and financial systems. AI tools support simulation, validation and regression testing to ensure accurate billing. Additionally, AI aids in designing billing catalogs, optimizing performance and developing interfaces, emphasizing the significance of modular design in building scalable COTS billing systems.

During a major media billing modernization project, normalized legacy billing logic was introduced, modular microservices were implemented and DevOps pipelines were established for continuous integration and deployment. AI tools like GitHub Copilot and Windserf were integrated into the development workflow to accelerate code generation and enhance quality. This initiative ensured business continuity and set the groundwork for future COTS integration by aligning the architecture with industry standards and enabling scalable, testable components.

Migrating to COTS products, particularly convergent billing systems, can significantly enhance long-term maintainability and scalability. But success depends on rigorous upfront rationalization, clear mapping of old-to-new processes and robust fallback strategies during cutover phases.

Minimizing Business Disruption While Supporting Innovation

One of the most difficult balances to strike in modernization projects is minimizing disruption while enabling innovation. It’s not just about ripping out old systems; it’s about gradually introducing new capabilities without alienating users or destabilizing operations.

Progressive modernization, starting with loosely coupled modules like reporting, analytics, or self-service portals, can help prove out new technologies before moving into core transactional systems. During this time, digital twins or shadow deployments can simulate new architectures alongside legacy systems without impacting production workloads.

Progressive modernization with loosely coupled modules like reporting, analytics, or self-service portals allows for demonstrating new technologies before full integration. Digital twins or shadow deployments can simulate new architectures without affecting production.

A phased strategy was used in the media billing modernization program to minimize disruption. Non-critical components, such as reporting and invoice generation, were decoupled from the legacy system and redesigned using microservices on AWS. A shadow deployment model allowed the new system to run in parallel with the old one, simulating real-world scenarios.

Regular workshops with business users gathered feedback and ensured alignment. Training sessions helped familiarize teams with new interfaces and workflows. By the time core billing logic was transitioned, the organization was accustomed to the new architecture, reducing resistance.

This approach ensured a smooth transition and built confidence in the modernization roadmap.

Modernization efforts must also consider regulatory compliance, particularly when handling customer data. Adopting zero-downtime deployment strategies and embedding robust rollback mechanisms into DevOps workflows can help organizations deliver faster without compromising stability or trust.

Future-Proofing With Big Data, API Architecture and Convergent Billing

To remain competitive, telecom providers must move away from monolithic systems and embrace modular, interoperable platforms. This means building infrastructure that’s not only agile, but also extensible—ready to adapt to the next wave of customer demands, partner integrations, or pricing models.

Big data plays a central role here. With the right data ingestion, storage and processing frameworks in place, providers can generate real-time insights on network performance, customer behavior and fraud detection. AI models trained on this data can recommend optimizations in areas ranging from pricing to infrastructure planning.

Equally important is the shift to API-first architecture. APIs enable seamless integration across platforms and make it easier to experiment with new services, partners, or monetization models without needing to overhaul core systems.

Convergent billing systems, which combine prepaid, postpaid and multi-service offerings, are also key to building a unified view of the customer and reducing the complexity of managing multiple platforms. These systems make it easier to launch bundled offers, track usage across products and reconcile billing more efficiently.

Telecom modernization is no longer a one-time transformation. it’s a continuous evolution. With generative AI to accelerate development, DevOps to ensure safe and repeatable deployments and modular architecture to enable flexibility, providers can break free from the constraints of legacy systems without jeopardizing performance or customer trust.

Telecom modernization is ongoing. Generative AI, DevOps and modular architecture enable us to move beyond legacy systems while maintaining performance and trust. AI speeds up development, DevOps ensures safe deployments and modular design provides agility. Tasks like analyzing billing rules or validating system migrations now take days instead of months. Telecom platforms will become self-optimizing with embedded AI agents, transforming from finite projects to adaptive capabilities.

Telecom modernization is a continuous evolution. With 24 years of experience in global digital transformations, I’ve seen how legacy systems can be revamped using COTS platforms, modular architecture and technologies like AI.

GitHub Copilot boosts development with DevOps pipeline automation. AI agents will continually optimize telecom platforms, making transformation adaptive.

The path forward is not without risk, but the cost of inaction is far greater. Organizations that approach modernization with a pragmatic, tech-forward mindset can expect to reduce operational overheads, launch new products faster and future-proof their infrastructure for the next wave of digital demand.

July 3, 2025
Break the Bottleneck of API Sprawl With AI-Powered Automation

The race to accelerate digital transformation across business units within an organization has led to a rapid surge in APIs, though unfortunately, without a central strategy in place.

In 2023, the majority of internet traffic (71%) was API calls, with typical enterprise sites seeing an average of 1.5 billion API calls. Large companies with over 10,000 employees tend to have over 250 internal APIs.

This challenge is exacerbated by the growth of SaaS applications in organizations’ IT landscapes, with each application bringing its own integration and API needs. On average, an organization manages over 1,000 different apps. Besides, it is not just SaaS — most organizations possess a large repository of apps and APIs across their diverse hybrid landscape, and more often than not, do not possess a single dashboard to view and manage them all.

Decentralized development teams operating without visibility or governance often create duplicate APIs and fail to retire outdated ones, further intensifying API sprawl and system clutter.

The result: Growing development, maintenance and operational costs, reduced developer productivity and increased security and compliance risks.

Most organizations have multiple vendor solutions in place, including siloed integration middleware installations, home-grown point-to-point integration solutions and API management offerings. These pile on to the challenges posed by API sprawl. Lack of rationalized and streamlined integrations across application landscapes makes it extremely difficult to drive innovation, transformation and growth in such companies.

The Opportunity: To Tame API Sprawl and Create Seamless Connectivity

Companies that can tame API sprawl and master seamless connectivity and integration across systems, apps, events and data and across hybrid, multi-cloud environments will be the ones to unlock faster innovation and sustainable growth.

For instance, a customer service representative at a large retailer receives a call from a frustrated shopper whose package hasn’t arrived as expected. The representative searches multiple systems, including order tracking, email communications and inventory management and finds conflicting information on all accounts. This lack of connected data will not only result in delayed issue resolution but also erode trust in the customer experience.

Meanwhile, behind the scenes, the issue stems from disconnected systems, fragmented APIs and siloed data that can’t easily be accessed or integrated. Some systems don’t talk to each other at all. Others require expensive, manual workarounds for connectivity. To top this, the retailer must comply with regional data privacy regulations, making it harder to share data appropriately across the business.

Now, imagine a solution that gives IT teams unified visibility into all integrations across cloud and on-prem systems, intelligently recommends existing APIs for reuse and automates secure, policy-compliant connections between applications.

With this level of seamless integration, the same customer inquiry could be resolved in seconds, not hours or days, by providing the customer service representative a single, unified view of all systems.

A Unified Integration Strategy

This solution to API sprawl and fragmented integration cannot be resolved by simply adding more technology. There are already too many silos of integration tools in the technology stack. The answer lies in a unified integration strategy led by AI-powered automation and the use of AI agents.

Autonomous AI agents can understand complex instructions to optimize workflows. They act as an intelligent integration layer by maximizing efficiency. They rapidly discover existing APIs, recommend optimal reuse of integrations to avoid duplication and automate the creation of new integrations to address various patterns of integration.

Now, complex tasks like stitching together data and systems across sales, inventory, fulfillment and customer service are simplified. AI agents are quick to define, deploy and manage APIs and integrations more autonomously, thereby drastically improving operational efficiency and customer experience.

For the large retailer mentioned above, AI and automation would be the answer to its end-to-end customer journey — from the moment an order is placed, to fulfillment, shipment and post-delivery engagement. Customer-facing digital applications are seamlessly integrated with inventory and supply chain systems, whether legacy or modern, and AI-powered automation ensures everything works in harmony.

Retail is just one example. We can see the approach coming to life in other industries as well. To truly transform, businesses must adopt a unified integration strategy, powered by AI, which connects front-end experiences with back-end operations across hybrid and multi-cloud environments.

In banking, third-party collaborators are securely integrated with core banking systems to support embedded finance and banking-as-a-service models, opening up new revenue streams. In energy and utilities, advanced connectivity and integration is critical for micro-grids, roof-top solar units and smart meters that offer real-time information on electricity, gas and water usage. Technologies like AI, IoT, edge computing and cloud help optimize grid operations, predict equipment failures, identify water leakages and deliver more reliable, transparent service to customers. 

The retail example comes full circle with logistics. These organizations can connect multi-modal shippers (sea, rail, road), freight forwarders and logistics hubs, all while maintaining continuous connectivity to e-commerce shipments for dynamic routing and real-time requests, enabling more flexible, reliable and faster service.

Overall, the impact is tangible: Modern integration strategies would drive innovation, reduce operational friction and empower both developers and business users to accelerate business success, in a bid to improve customer experience.

April 25, 2025
Can AI Replace DevOps Engineers ?

There were system administrators. They conducted backups, designed VPNs and created disaster recovery plans (DRPs). Managers didn’t understand them and thought they were barely working. For a while, everything was calm and good.

Then came DevOps. The development immediately sped up, pipelines started working better, products had fewer bugs and the developers were finally happy. For a while, everything was good again.

Then the containers arrived. Everyone rushed into Kubernetes as if the train had already left the station. They practiced Kubernetes for the sake of Kubernetes, for the buzzword, and the chance to show they were at the cutting edge. For a while, everything was good again.

And then artificial intelligence (AI) came — a revolution, no less. Hysteria ensued in the IT sector about how nobody would be needed in the foreseeable future as everything would be done by AI. But before we start panicking, let’s figure out what AI does well, where it falls short in DevOps, and whether DevOps professionals should fear it and start re-skilling today.

Where AI Excels and Where It Stumbles

AI is a superb helper in automating repetitive tasks and an excellent assistant in development. It checks syntax, makes code beautiful, inserts functions, writes code comments, finds bugs you might have missed, succeeds in testing and generating tests and performs well in log analysis. It also helps predict system behavior, loads, and issues and works well in detecting threats and responding to events.

However, AI lacks a deep understanding of context, critical analysis, creativity and motivation — crucial aspects of DevOps, which is fundamentally about communications. AI also struggles with building complex processes and interactions between diverse and intricate parts of a system.

Where DevOps outshines AI

DevOps is a practice that promotes collaboration between software developers and IT professionals while automating software delivery and infrastructure changes. Its goal is to create a culture and environment where development, testing, and releasing software can happen rapidly, frequently and reliably. AI is not yet the magic wand that can solve human-centric tasks.

DevOps extends beyond pipelines, servers, Kubernetes and Docker, emphasizing teamwork and strong relations within a team. With DevOps, no separate teams are sticking to their specific rules; instead, there is a unified delivery team where everyone is responsible for issues and eager to solve them. DevOps also involves inter-departmental communication, knowledge transfer, and interaction during deployments.

What About the Technical Side?

Naturally, AI can assist you in writing scripts in Bash, Python or PowerShell — ask ChatGPT and it will be provided along with comments. However, thinking through a complete DRP with testing and backup verification is beyond AI capabilities. How do you explain to it that the tapes we use for backups need to be stored in a bank and the data recovery process should be tested quarterly, potentially failing due to a miswritten phone number?

Or for instance, take containers and Kubernetes. Precise configurations of your deployments require a lot of work — PSP, PodDisruptionBudget, good limits and requests, solid probes (set up so they don’t crash the cluster), security configuration and proper monitoring. Relying on AI for these tasks is impractical — you need a DevOps engineer competent to implement them.

AI vs. DevOps: Who is the Winner?

How often have you written a script and felt too lazy to add comments or write a README? How many times have you manually searched for a bug in the code and fixed it (line by line, through long nights)? How frequently have you had to wake up at night due to a production notification? This is where AI comes in, automating those and similar tasks.

There will be no battle for survival in the future. The strength of AI is unleashed when combined with a human-centric approach. That means, DevOps will plan, execute, control, and improve, while AI will help handle tasks faster and with greater accuracy.

Good Specialists Learn Continuously

DevOps must improve their knowledge, hone skills and learn to handle conflict situations — not just because of perceived AI threats, but to become and remain in-demand specialists. If you are flexible, easily adapt to changes and face challenges head-on without lamenting over the past, no AI can replace you. AI will assist you, asking, “How’s it going today? What can we improve and accomplish today?”

June 18, 2024
Best of 2023: ‘Scrum == Cancer’ ¦ Plus: Linux 6.5 Ships

As we close out 2023, we at devopsy.kinsta.cloud wanted to highlight the most popular articles of the year. Following is the latest in our series of the Best of 2023.

Welcome to The Long View—where we peruse the news of the week and strip it to the essentials. Let’s work out what really matters.

This week: Scrum sucks, sources say; and here comes the Linux 6.5 kernel. (more…)

December 29, 2023
Microsoft kills Python 3.7 ¦ … and VBScript ¦ Exascaling ARM on Jupiter

Welcome to The Long View—where we peruse the news of the week and strip it to the essentials. Let’s work out what really matters.

This week: VS Code drops support for Python 3.7, Windows drops VBScript, and Europe plans the fastest ARM supercomputer. (more…)

October 11, 2023
Raspberry Pi 5: Faster, Better, Stronger — Spendier

Welcome to The Long View—where we peruse the news of the week and strip it to the essentials. Let’s work out what really matters.

In a cheeky extra post this week: Everyone’s favorite single-board ARM computer, the Raspberry Pi, has a new generation coming soon. Compared to the ’4, RPi5 has double the performance, quadruple the base RAM and far more capable I/O. (more…)

September 29, 2023
FCC: Net Neutrality is a ‘Thing’ Again ¦ Meta Shutters Big UK Site

Welcome to The Long View—where we peruse the news of the week and strip it to the essentials. Let’s work out what really matters.

This week: Network neutrality is back on the FCC’s agenda, and unused London office gets Zucked. (more…)

September 27, 2023