Mikhail Shilkov is a software engineer and cloud infrastructure expert with deep experience in AI-assisted developer tools, Infrastructure as Code, and Pulumi.
54 articles from this blog
A tutorial on verifying and load-testing Azure Functions using the Loader.io tool, covering two verification methods.
Explains how the Azure CLI securely stores and manages OAuth access tokens locally to authenticate users between command executions.
A tutorial on building a globally-distributed URL shortener using serverless Azure Functions and Cosmos DB, defined entirely in a Pulumi TypeScript program.
A tutorial on using Pulumi to deploy a static website on Azure, covering storage, CDN, and TLS setup.
A tutorial on deploying serverless Azure Functions using Pulumi's callback syntax, simplifying infrastructure and code deployment.
A guide to using Pulumi's infrastructure-as-code to deploy and manage Azure PaaS applications like App Service and SQL Database.
Analysis of Google Cloud Functions' improved cold start behavior, showing less frequent recycling of idle instances compared to AWS and Azure.
A technical analysis demonstrating how Binaris serverless functions eliminate cold starts, with visual comparisons to AWS, Google, and Azure.
Analysis of cold start durations in Azure Functions, comparing deployment methods and the impact of Application Insights on performance.
Explores concurrency and isolation models in AWS Lambda and other serverless platforms, analyzing trade-offs for different workloads.
A visual experiment comparing cold start performance of serverless functions across AWS Lambda, Google Cloud Functions, and Azure Functions.
Author discusses limitations of static blog posts for serverless performance reviews and proposes an automated, regularly updated website solution.
Argues for using TypeScript over YAML for cloud infrastructure automation, using a serverless URL shortener on AWS as a case study.
A technical analysis of serverless HTTP function scalability, testing performance under high traffic loads comparable to StackOverflow across major cloud providers.
