April 2018 Summaries

The author of the article is leading the distributed tracing project at New Relic and recognizes that the term "tracing" can refer to different concepts in the industry. The project spans various domains, including data collection, propagation, and display, which are related but distinct. The author attended a workshop where Ben Sigelman presented a framework for understanding these distinctions, highlighting the differences between analyzing transactions, recording transactions, and describing transactions. New Relic is joining the OpenTracing standard to provide better instrumentation support and make it easier for customers to choose their monitoring solution, with the goal of increasing visibility and diagnosing problems quickly. By defining clear terms and problem domains, the industry can reduce confusion and reason better about how different standards solve specific problems and work together.

Distributed tracing, as explored by Erika Arnold at New Relic, involves analyzing, recording, and describing system transactions to enhance performance monitoring and troubleshooting. New Relic's distributed tracing project encompasses data collection, propagation, and visualization, spanning multiple domains within the company. The project aligns with emerging standards like OpenTracing and TraceContext, aiming to improve context propagation across systems to prevent trace breaks. By adopting these standards, New Relic seeks to enhance its tracing capabilities, enabling seamless tracking across diverse microservices environments and facilitating better diagnostic insights for users. The article also emphasizes the importance of well-instrumented applications for efficient problem-solving and the potential of standardized APIs to reduce switching costs and encourage widespread adoption.

Modern software teams are increasingly adopting cloud-based solutions for hosting and deploying applications and infrastructure. However, deploying applications in the cloud can be complex and requires automation to streamline processes. To address this challenge, many teams are embracing continuous integration and continuous deployment (CI/CD) practices, which involve automating tasks from build to package to release. Various tools specialize in helping with these tasks, such as AppVeyor for Windows and .NET applications, AWS CodeDeploy for Amazon Web Services customers, and Kubernetes for container orchestration. Other notable tools include Bamboo, CircleCI, Codeship, Google App Engine, Helm, Heroku, IBM Cloud, Jenkins, Octopus Deploy, Red Hat OpenShift, Travis CI, Up, and Visual Studio Team Services (VSTS). Each tool has its strengths, weaknesses, and learning curves, and teams should carefully evaluate their needs before selecting a solution.

Modern software teams are increasingly opting for cloud-based solutions over on-premise ones, leading to a shift in how applications are deployed, with particular emphasis on automating the deployment process through continuous integration and continuous deployment (CI/CD) practices. To aid in this transition, a variety of tools have emerged, each catering to different needs and preferences, such as AppVeyor for Windows and .NET developers, AWS CodeDeploy and AWS Fargate for Amazon Web Services users, and Google App Engine for those on the Google Cloud Platform. These tools offer diverse features, from integrating seamlessly with popular developer tools to providing extensive language support and facilitating containerized deployments with platforms like Kubernetes and Docker. While solutions from major cloud providers like AWS, Google Cloud, and Azure offer comprehensive benefits, other platforms like Jenkins or CircleCI are favored for their flexibility and integration capabilities. The blog suggests that these tools, with their various strengths and learning curves, should be chosen based on the specific requirements of a team's cloud adoption strategy to optimize performance and customer experience.

Gannett's infrastructure is managed using containers and Kubernetes, but without proper monitoring, teams can lose track of what's happening in their clusters, leading to issues like capacity problems, container crashes, and API unresponsiveness. New Relic's Infrastructure on-host integration for Kubernetes provides deep monitoring of the container orchestration layer, collecting metrics on nodes, Namespaces, Deployments, ReplicaSets, Pods, and containers to provide total visibility, alerting, and dashboards for all Kubernetes entities. This integration helps teams monitor user experience, applications, containers, deployments/pods, nodes, cluster capacity, and resource utilization, allowing them to troubleshoot issues quickly and efficiently. With this integration, teams can move from an infrastructure-centric view to an application-centric one, combining application metrics with Kubernetes metrics to gain a deeper understanding of application performance in their clusters.

Gannett, a media giant, utilizes Kubernetes for its containerized infrastructure but faces challenges in maintaining visibility and control due to the automated nature of orchestration. To address these issues, New Relic has introduced an Infrastructure on-host integration for Kubernetes that provides comprehensive monitoring of container orchestration layers. This integration allows teams to track metrics related to nodes, Namespaces, Deployments, ReplicaSets, Pods, and containers, facilitating full visibility into both frontend and backend applications within Kubernetes clusters. The integration offers dashboards and alerting capabilities, enabling users to transition from an infrastructure-centric to an application-centric view, thereby improving the ability to troubleshoot and optimize performance. By combining application and Kubernetes metrics, New Relic aids in understanding application performance and infrastructure challenges, making it easier for teams to identify root causes of errors and improve deployment efficiency. The integration is available in public beta to New Relic Infrastructure customers and supports platforms such as AWS, Microsoft Azure, Google Cloud Platform, and IBM Cloud Container Service.

Many companies are adopting cloud computing as a core component of their technology strategy, but they often underestimate the costs associated with cloud migration. To accurately estimate these costs, organizations need to consider three main categories: current application and infrastructure costs, qualitative financial estimates of cloud infrastructure costs, and cloud migration costs. The latter includes the cost of migrating systems and data to the cloud, as well as labor costs for porting apps and creating inventory. A financial plan for a cloud migration requires building a timeline, understanding labor resources, and tracking weekly operating costs for both environments. Additionally, organizations need to benchmark their target environment to ensure operational parity with benchmarks before retiring original systems.

As digital enterprises increasingly adopt cloud computing, they often encounter challenges in effectively planning and budgeting for cloud migration due to underestimated costs and lack of financial expertise. Many organizations mistakenly view cloud migration as a simple "lift and shift" process, overlooking the complexities highlighted by a Forrester Consulting study that revealed widespread "cost complexity" and misaligned expectations. To mitigate these issues, companies must consider three main cost categories: current application and infrastructure costs, qualitative financial estimates of cloud infrastructure costs using calculators from major providers, and the often underestimated cloud migration costs, which include labor and potential use of migration automation software. Tracking and benchmarking during and after the migration is critical to avoid running parallel environments longer than necessary, ensuring a smoother transition and operational parity between the original and cloud systems. This strategic approach helps in managing migration timelines and costs effectively, ultimately leading to successful cloud adoption.

Cloud migrations require automation and configuration of new cloud infrastructure to reduce costs and free up resources for mission-critical innovation. Various tools are available to help with this process, including AWS CloudFormation, Puppet, Ansible, Chef, Kubernetes, Terraform, Google Cloud Deployment Manager, Microsoft Azure Automation, Cisco Intelligent Automation for Cloud, SaltStack, and VMware vCenter Configuration Manager (VCM). Each tool has its strengths, weaknesses, and learning curves, and the best choice depends on the specific needs of the project and teams involved. Combining two or more tools may be necessary to achieve maximum value from cloud infrastructure.

Cloud migrations offer significant benefits such as reduced toil and operating costs, with automation playing a crucial role in configuring new cloud infrastructures. Various tools are available to facilitate this process, including AWS CloudFormation, Puppet, Ansible, Chef, Kubernetes, Terraform, and more, each with unique features and integrations. These tools help automate the provisioning, configuration, and management of cloud resources across public, private, and hybrid environments, allowing organizations to focus more on innovation rather than routine maintenance. While products like CloudFormation and Google Cloud Deployment Manager cater to specific cloud providers, others like Puppet, Chef, and Ansible offer broader cross-platform capabilities. Selecting the right tool depends on project needs, team capabilities, and specific cloud infrastructure requirements, with some scenarios benefiting from combining multiple tools. The blog emphasizes that effective tool choice and integration can lead to successful cloud adoption and enhanced infrastructure management.

Superman vs. Batman and Coke vs. Pepsi are comparisons of rival brands, but Java and Python are programming languages with different use cases and fan bases. Despite their differences, both languages share similarities such as powerful communities, a large array of libraries, and the ability to convey tasks to machines in clear terms. However, they also have distinct differences, including compilation vs. interpretation, syntax, and threading models. Java is more popular, but Python's growth has been astronomical, especially in developed countries, due to its developer productivity, flexibility, and ease of learning. Both languages have their strengths and weaknesses, with performance being a complex metric that depends on environment, coding style, and libraries used. Ultimately, the choice between Java and Python depends on meeting a developer's requirements for conveying tasks to machines in the most straightforward manner possible, considering factors such as productivity, flexibility, and performance.

Java and Python are two of the most popular programming languages, each with unique strengths and weaknesses, as well as distinct use cases that cater to different developer preferences and needs. While Java is a compiled language known for its performance in creating large-scale, cross-platform applications, Python, being an interpreted language, excels in areas like data science, artificial intelligence, and machine learning due to its readability and flexibility. Both languages boast a vast array of libraries and strong community support, making them versatile tools for developers. Java's reputation has been marred by security issues, particularly with its browser plug-in, whereas Python is favored for its ease of learning and use, despite its own security challenges. The choice between these languages often comes down to the specifics of a project, the developer's expertise, and the requirements for productivity, speed, and clarity. As programming continues to evolve, understanding multiple languages, including Java and Python, can provide developers with the flexibility to tackle a wide range of projects and challenges.

DevOps is a combination of software development and operations philosophies, practices, and tools that aims to bring traditionally siloed teams into greater alignment. While it's often associated with technology, DevOps is ultimately a matter of culture and involves people, processes, and tools. The term "DevOps" was coined by Patrick Debois and Andrew Shafer in 2008, but the concept has evolved over time to become a widely adopted practice across various organizations. Despite its growing popularity, many people still don't fully understand DevOps or how it works, and even seasoned veterans may retain blind spots. However, DevOps is not just about technology; it's also about culture, measurement, and business success. It can be applied beyond IT teams to improve overall organizational efficiency and quality, making it a powerful business strategy as well as a tactical approach.

DevOps, a fusion of software development and operations, is a cultural shift rather than a purely technological one, emphasizing collaboration and continuous improvement across teams. Originating from a 2008 Agile conference, the term highlights the integration of traditionally siloed development and operations teams, yet its practice involves a broader cultural change that resists completion or isolation. Despite debates over job titles like "DevOps Engineer," the role is prevalent with high compensation, reflecting its growing importance in achieving business success through improved IT performance and market agility. DevOps principles—speed, agility, and collaboration—are being applied beyond IT departments to enhance business strategies, and its value can be quantified through metrics like deployment frequency and mean time to recovery, which are crucial for organizational efficiency and competitive advantage. The approach is not limited to small tech companies but is increasingly adopted by large enterprises, demonstrating its universal applicability and correlation with business success.

New Relic is a proactive technology administration strategy that helps companies move beyond the classic break-fix IT model by leveraging advanced tools to monitor and diagnose system performance issues against expected outcomes. This approach can have significant and quantifiable results in areas such as total cost of ownership, human resources productivity, infrastructure efficiencies, e-commerce revenues, subscription revenues, advertising revenues, and mobile app adoption. By identifying and addressing system-performance problems before they impact the customer experience and bottom line, companies can improve their overall system performance, reduce errors and associated subscriber attrition, and boost efficiency to drive productivity, ultimately gaining a competitive advantage in their digital ecosystem.

In today's technology-driven business environment, every company grapples with the challenges of operating efficiently, maximizing revenue from digital transformation, and identifying new opportunities. To address these, businesses are encouraged to transition from a traditional break-fix IT model to a proactive technology strategy, which involves anticipating and resolving system issues before they affect customer experience and revenue. New Relic supports this approach by offering tools for comprehensive monitoring and diagnostics that enhance IT ecosystems' predictability and performance. By implementing New Relic's multi-tenant solutions, companies can achieve significant cost savings, boost productivity, and enhance revenue streams across various sectors, including e-commerce, subscriptions, and advertising. This proactive strategy not only provides a safeguard against potential system failures but also serves as a strategic asset that can improve competitiveness and operational efficiency.