What is a potential disadvantage of Serverless architecture?

What is a potential disadvantage of Serverless architecture?

• A. Latency during periods of low activity
• B. Simpler cost model with no hidden costs
• C. Guaranteed immediate cold start
• D. No vendor lock-in

Answer: (A)

Serverless functions are not always running; they’re started on demand. When there hasn’t been recent activity, the platform may remove or suspend the execution environment, so the next invocation has to spin up a new container, load the runtime, and initialize the function code. This startup delay, known as a cold start, creates latency. In workloads with sporadic or low activity, these cold starts can occur more often or be more noticeable, making response times less predictable. That latent startup cost is the key disadvantage of serverless architectures. The other ideas don’t fit as well. Costs in serverless can be unpredictable or have hidden aspects, rather than being a simpler, all-clear model. Cold starts don’t guarantee immediate latency—quite the opposite. And vendor lock-in can actually be a concern, since your functions often rely on platform-specific services and APIs.

Serverless functions are not always running; they’re started on demand. When there hasn’t been recent activity, the platform may remove or suspend the execution environment, so the next invocation has to spin up a new container, load the runtime, and initialize the function code. This startup delay, known as a cold start, creates latency. In workloads with sporadic or low activity, these cold starts can occur more often or be more noticeable, making response times less predictable. That latent startup cost is the key disadvantage of serverless architectures.

The other ideas don’t fit as well. Costs in serverless can be unpredictable or have hidden aspects, rather than being a simpler, all-clear model. Cold starts don’t guarantee immediate latency—quite the opposite. And vendor lock-in can actually be a concern, since your functions often rely on platform-specific services and APIs.