Reserved concurrency limits: does the account pool auto‑scale?

Answer

When a function’s reserved concurrency is set higher than the current account‑level concurrency limit, AWS does not automatically increase the account pool. The reservation value you set remains unchanged, but the function’s actual concurrent executions are throttled to the account limit.

Confirmed facts

• Reserved concurrency is a static limit you configure per function; it does not auto‑scale based on traffic.
• The account‑level concurrency limit (default 1,000) is also static unless you request a limit increase through AWS Support.
• No documented mechanism exists that adjusts the account limit in response to reserved concurrency settings.
• Provisioned concurrency behaves similarly: it is a static baseline capacity that also does not auto‑scale.

Likely explanation

The reservation you configure is stored in the function’s configuration, but the effective concurrency available to that function is the lesser of:

1. The reserved concurrency value you set, and
2. The account‑level limit minus any concurrency already used by other functions (the unreserved pool).
3. If provisioned concurrency is also enabled, it adds to the reserved amount, but the sum still cannot exceed the account limit.

Consequently, if the reservation exceeds the limit, the function sees throttling (429 errors) once traffic tries to use more concurrent executions than the account allows, while the reservation value you set remains visible in the console and CLI.

Steps to verify and address the situation

1. Check the current account limit and usage:

   aws lambda get-account-settings

   Look for TotalConcurrentExecutionsLimit and UnreservedConcurrentExecutions.
2. Check the function’s reserved concurrency:

   aws lambda get-function-configuration --function-name <your-function-name>

   The ReservedConcurrentExecutions field shows the value you set (or is absent if none).
3. If the reserved value is greater than the account limit and you are seeing throttling, request a limit increase via AWS Support (or reduce the reservation).
4. Monitor throttling with CloudWatch metric Throttles for the Lambda function to confirm the behavior.

Missing diagnostic detail

To decide whether a limit increase is truly needed, please provide the current UnreservedConcurrentExecutions value from get-account-settings. If the unreserved pool is already exhausted, raising the limit is the appropriate next step; otherwise, lowering the reservation may resolve the throttling.