Serverless
SUR-ver-less
A cloud model where the provider manages all infrastructure and you only pay for actual execution time.
Serverless is a cloud computing model where you write functions, upload them, and the cloud provider handles everything else: provisioning servers, scaling, patching, and monitoring. You pay only for the time your code actually runs. No traffic, no cost.
AWS Lambda is the most well-known serverless platform. You upload a function, configure a trigger (an HTTP request, a database change, a scheduled event), and Lambda runs your function when the trigger fires. It scales from zero to thousands of concurrent executions automatically.
Serverless does not mean there are no servers. It means you do not think about servers. The provider manages them. This is liberating for small teams that do not want to manage infrastructure. But it comes with constraints: execution time limits, cold starts (the first request after inactivity is slower), and vendor lock-in.
Examples
A startup builds its API on Lambda.
Each API endpoint is a Lambda function. The API Gateway routes requests to the appropriate function. At 3am with no users, zero functions are running and the cost is zero. During a traffic spike, Lambda runs 500 concurrent instances automatically. The team manages zero servers.
Serverless handles a background job.
When a user uploads an image, a Lambda function triggers automatically: it resizes the image into three sizes (thumbnail, medium, full), stores them in S3, and updates the database. The function runs for 2 seconds and costs $0.00003. No dedicated image processing server needed.
Cold starts affect user experience.
The first request after 15 minutes of inactivity takes 3 seconds because Lambda needs to initialize a new execution environment. Subsequent requests take 50ms. The team implements provisioned concurrency (keeping functions warm) for customer-facing endpoints, accepting the cost for consistent performance.
In practice
Read more on the blog
Frequently asked questions
When should I use serverless vs. containers?
Serverless for event-driven workloads, APIs with variable traffic, and background jobs. Containers for long-running processes, workloads that need consistent performance, and applications that exceed serverless limits (execution time, memory). Many teams use both: serverless for APIs and containers for workers.
Is serverless cheaper than running servers?
At low to moderate traffic, yes. You pay only for execution time. At high sustained traffic, running your own containers is cheaper because serverless per-request costs add up. The crossover point varies, but most teams save money with serverless until they reach millions of daily requests.
Related terms
On-demand computing infrastructure (servers, storage, networking) provided by a third party.
Lightweight, portable packages that bundle application code with all its dependencies.
Application programming interface: a defined way for software programs to communicate with each other.
Running code on servers located close to users around the world instead of in a single data center.
Want the complete playbook?
Picks and Shovels is the definitive guide to developer marketing. Amazon #1 bestseller with practical strategies from 30 years of marketing to developers.