ราคาบริการแบบไร้เซิร์ฟเวอร์และระดับใช้งานฟรี

With serverless you pay per request and per compute time, never for idle servers. When nobody calls your function, you pay nothing.

Lambda cost has two parts: Requests (a flat price per million invocations) and Duration (billed per millisecond, scaled by memory).

Duration uses GB-seconds: memory in GB times run time in seconds. More memory or longer runs cost more. Here is a quick estimator.

The Lambda free tier gives 1 million requests and 400,000 GB-seconds every month — and it never expires after 12 months.

A 128 MB function running 100 ms can serve millions of requests inside the free tier, making Lambda perfect for hobby projects and prototypes.

More memory also adds CPU, so sometimes more memory finishes faster and lowers total GB-seconds and cost. Always test — never assume.

Unlike an always-on EC2 box, a deployed Lambda nobody calls costs zero compute. You only pay to store the code.

Lambda is cheap, but it triggers paid services: API Gateway requests, CloudWatch Logs storage, and data transfer all add to the bill.

Use the AWS Pricing Calculator to model monthly cost from invocations, average duration, and memory. Estimating early avoids surprises.

Set a billing alarm in CloudWatch or AWS Budgets so you get notified the moment spending crosses a threshold you choose.

For spiky or low-volume work, serverless usually beats an idle server. For steady, very high constant load, a reserved server can be cheaper.

Recap: Lambda costs per-request plus GB-seconds, has a big always-on free tier, trades memory for speed, and rewards estimating and alarms.