May 2022 Summaries

AWS offers a variety of pricing plans to suit different organizational needs, including pay-as-you-go, reserved instances, AWS spot instances, savings plans, and the free tier. The pay-as-you-go plan provides flexibility but can be expensive, while reserved instances offer significant cost savings for consistent usage. AWS spot instances allow organizations to purchase unused resources at a 90% discount, but with limited access. Savings plans enable customers to commit to a minimum spending amount and usage level for a set period, resulting in lower hourly rates for computing resources. The free tier offers three sub-tiers: always free, 12 months free, and trials, allowing organizations to experiment with AWS services without commitment. Using the AWS pricing calculator or optimizing cloud costs through automated solutions can also help reduce expenses. Understanding the various plans and resource limits is crucial to selecting the right plan for an organization and maximizing its existing resources.

Matas Kulkovas, a brilliant engineer, has developed an open-source security chaos engineering tool called Kirvis for Kubernetes, which helps handle security uncertainties at the cluster level by simulating various types of threats and testing how the system responds to them. The approach is based on a circular sequence of actions that involve hypothesizing how the system will behave under a given threat, inserting a security error, and verifying whether the system behaves as expected. Kirvis is the first SCE tool for Kubernetes and provides a more realistic understanding of how well certain practices perform under specific conditions, enabling continuous recalibration of security activities.

Google Cloud Platform offers a range of cloud computing services, including storage, compute, and managed Kubernetes. The platform has three main pricing models: pay-as-you-go, free tier, and committed use discounts. Pay-as-you-go is ideal for organizations with irregular usage, offering flexibility but at a higher cost per hour. The free tier provides access to limited resources for free, while committed use discounts offer up to 70% off for long-term commitments. Spot VMs are available at lower prices but may be reclaimed by Google Cloud with short notice. New customers receive $300 in credits to try out the platform. The pricing calculator can help organizations estimate costs and choose the best plan for their needs, which is similar to other providers such as AWS and Microsoft Azure, although AWS may be more cost-effective for standard options.

The global economy is experiencing a shift as investors move towards safer equities and venture capitalists lose interest in cash-burning startups. Tech companies that saw significant user base increases during the pandemic are now facing challenges due to rising prices and reduced consumer spending on rides or meal delivery services. Additionally, tech giants like Apple and Tesla are struggling with production issues in China due to pandemic-related lockdowns. Businesses have an opportunity to increase their gross margins by reducing cloud costs, which can make a significant difference immediately. Despite this, many companies continue to struggle with balancing cost and performance in the cloud, often resulting in wasted resources and missed discount opportunities. Specialized automation tools like CAST AI can help optimize cloud infrastructure and reduce costs without requiring additional engineering effort or upfront costs.

The AWS Cost Management console is a feature that enables users to better understand, organize, and detect anomalies within their cloud billing, providing tools such as AWS Cost Explorer, AWS Budgets, cost anomaly detection, and Savings Plans to optimize costs without compromising on performance. By using these tools, users can gain a clearer understanding of their finances, forecast future expenses, set budgets and notifications, identify potential anomalies, and manage costs more efficiently. The console helps users reduce waste, find the perfect ratio between performance and cost, and save money without affecting cloud operations, ultimately gaining better control over their AWS spending.

Autoscaling is a significant challenge for teams managing Kubernetes clusters on Azure Kubernetes Service (AKS). The automated Advanced Autoscaler by CAST AI optimizes cluster performance and cost efficiency. It uses real-time autoscaling, headroom policy, and spot fallback mechanisms to adjust resources according to workload demands. Users can optimize their AKS clusters in less than 15 minutes with the Advanced Autoscaler, which identifies potential savings through rightsizing and better VM selection. The platform rebalances infrastructure and continuously optimizes the cluster using smart autoscaling mechanisms. CAST AI's settings allow users to customize how new nodes are added and managed.

The Russian-Ukrainian war has significantly impacted the world of fintech, with many companies adapting their services to uphold international sanctions and support humanitarian aid. This rapid response is made possible by lean IT infrastructure and cloud optimization, including Kubernetes automation. Automating cloud operations can streamline processes, reduce costs, improve security, and enable continuous integration and deployment (CI/CD), ultimately increasing a fintech's agility and resilience in the face of unexpected challenges.

AWS Cost Explorer is a popular tool for tracking usage reporting, but it has limitations when it comes to providing real-time cost visibility. The current billing system updates its data at least once a day, which can lead to unexpected costs and surprises. To overcome this, users need access to the raw data used to create the invoice, known as the Cost and Usage Report (CUR), which can be fed into the Cost Explorer for visualization. However, AWS only refreshes this data file three times a day at most, and there is no guarantee of updates beyond that. Real-time visibility into cloud costs is essential to avoid massive cost spikes and optimize regular cloud bills, but it's just one side of the coin - doing something about it with automation is also necessary for effective cost optimization tactics such as autoscaling, rightsizing, spot instance use, and more.

Google Cloud Platform has introduced significant price increases across various core services around storage, indicating that the public cloud is no longer expected to continue getting cheaper. The high inflation rates and technology shortages are likely to affect the cloud industry's pricing strategy. Cloud service providers like AWS have reduced prices a total of 107 times since its launch in 2006, but these regular price cuts were not sustainable due to the high gross margins enjoyed by the providers. With the current inflation rate and supply chain troubles, it is unlikely that cloud providers will continue to lower their prices. Instead, companies should prepare for potential price hikes by reserving cloud resources while they're cheap, optimizing existing resources, and utilizing automated cost optimization tools. These tools can help select the right types and sizes of cloud resources, scale resources up and down automatically, decommission unused resources, and automate the use of spot instances to achieve significant cost savings.

Explain the WHY behind cost optimization is key to making developers care about it. Cloud operations teams can't rely on just implementing analysis/optimization tools and expecting engineers to implement changes without a clear narrative on the value of cost optimization. A clear explanation of why optimizing cloud costs is important, such as removing waste via smart money management and good hygiene practices, helps devs understand the importance of FinOps. Building visibility into cost drivers with tagging allows teams to allocate costs to specific resources, making it easier for engineers to make informed decisions about cost-effectiveness. Investing in real-time monitoring and alert mechanisms ensures that teams can catch and prevent costly mistakes, such as running an instance longer than needed or failing to turn off a job. Implementing automation solutions takes these tasks off the plate of individual engineers, allowing them to focus on their core work without having to worry about cloud expenses.

Fintech companies face significant challenges in predicting and managing their operational expenses (opex) due to the dynamic nature of the industry, leading to constant battles for profit margins. Cloud costs are a major contributor to this struggle, as fintechs often experience high barriers to planning, managing, and reporting on cloud costs throughout a project's lifespan and across multiple teams. The key to overcoming these challenges lies in implementing best practices such as building a culture of cloud cost ownership, avoiding cloud waste with automated rightsizing and autoscaling, preventing cloud sprawl with automated resource scheduling, not falling into the trap of reservations or savings plans, and utilizing spot instance automation to cut costs. By adopting these strategies, fintech companies can improve their ability to scale while staying cost-effective, even in the face of unexpected events such as sudden increases in demand or changes in market conditions.

Cloud computing has revolutionized software engineering for businesses, offering new ways to complete tasks and manage global deployments of updates and new software efficiently. However, one major hurdle is cloud cost control, with 70% of US businesses and IT departments facing challenges in managing costs, which can exceed expectations by as much as 62%. To achieve effective cloud cost management, companies must understand where costs are coming from, who is using resources, and estimate future needs through activities such as reporting, visibility, cost monitoring, forecasting, allocating, and budgeting. The primary goal is to gain control over cloud costs through careful management, which requires understanding how much, where, and why usage occurs. Cloud automation is a crucial component of cloud optimization, enabling the use of instance rightsizing, automated resource scaling, spot instance use, and removing unused resources, making it easier to manage cloud costs without manual work. By leveraging cloud automation, companies can achieve tangible financial results, such as increased gross margin, and provide the best user experience possible while being cost-efficient.

Kubernetes and Docker are two separate technologies that can complement each other when used correctly. Kubernetes provides an API to orchestrate containers, simplifying the development process, optimizing IT costs, and offering cloud migration options. It allows for decreased development timeframes, optimized IT costs, and cloud migration options. On the other hand, Docker enables developers to build, share, and run various apps, providing versatility, a consistent isolated environment, and easy collaboration. Although they are both open-source cloud-native technologies, Kubernetes is designed to run in a cluster using multiple virtual machines, whereas Docker focuses on packaging containerized applications using a single node. Despite their differences, both can be used separately or together, with the latter being a popular choice for many enterprises serving microservices development needs.