October 2021 Summaries

AWS Instance Purchasing Options offer flexibility in computing power with On-Demand, Savings Plans, Reserved Instances, and Spot Instances. The most flexible yet expensive plan is On-Demand, suitable for startups and companies meeting part-time or sudden performance spikes. Savings Plans provide up to 72% savings by committing to a specific instance family and region, while Reserved Instances offer higher savings but require more management. Spot Instances provide the highest savings but come with high risk. Companies can adopt various payment models to suit their computing needs, such as using Standard RIs for fixed computing needs or Compute Savings Plans for fluctuating usage. By choosing the right plan, companies can reduce infrastructure costs and optimize their cloud spending.

Containerization has improved application development, agility, deployments, portability, operational efficiency, scalability, modernization, and life cycle management for enterprises. However, allocating costs to containers in environments with multiple containers, projects, and teams is a significant challenge. Cloud service providers charge for server instances that comprise clusters, and traditional billing for managed Kubernetes is per cluster per hour plus additional underlying resources consumed. Amazon EKS charges per hour for each cluster created, allowing multiple applications to run on a single cluster. Identifying how container costs are generated is part of understanding the allocation of costs to containers and Kubernetes for better cloud resource management.

The Archera team, formerly Reserved.ai, implemented a tool to accelerate their product feedback loop by improving collaboration between engineering and product teams. They researched various tools, including local development environments, collaborative development platforms, Kubernetes-specific solutions, and finally settled on Tilt, which offered quick and simple setup, clear documentation, native support for Kubernetes + Helm, live update capabilities, a UI for tracking build progress, reproducible snapshots, and aligns with their open-source values. With Tilt, the team was able to shorten their feedback loop, enabling rapid prototyping and usability experiments, and ultimately achieved their launch date ahead of schedule while being prepared to adjust during launch. The team is now integrating Tilt into their process and exploring additional capabilities to streamline local dev environments.

KubeCon is a Kubernetes conference that originated at Google as Borg project, later open-sourced in 2014, revolutionizing cloud computing. The first KubeCon was held in 2015, focusing on topics directly related to Kubernetes. Over time, its scope expanded to encompass broader cloud-native technology discussions, now hosted by the Linux Foundation's Cloud Native Computing Foundation. This year's conference will feature twelve tracks covering various aspects of cloud computing, with talks from renowned companies and experts, aiming to promote engagement and education in open-source and cloud-native technologies.