AWS Lambda Returns TooManyRequestsException After a Reserved Concurrency Change
Overview
AWS Lambda returns TooManyRequestsException (HTTP 429) when an invocation would exceed the function's reserved concurrency or the account's Regional concurrency limit; a reservation of 0 throttles every call. Capture the reservation and account settings, find the change in CloudTrail, then restore or resize the reservation while leaving 100 unreserved, unless the owner set 0 on purpose.
Diagnose
Symptom
What responders observe.
The usual trigger is a change to the function's reserved concurrency: a PutFunctionConcurrency call that set it to 0 or to a value below the function's normal peak, for example from infrastructure code or the console.
Reserved concurrency is both a floor and a ceiling.
It guarantees the function that much capacity and stops it scaling past it, and the setting applies to the function as a whole, including every published version and $LATEST.
The second trigger is account-wide.
Every reservation is carved out of the Regional account limit (1,000 concurrent executions by default, which can be raised), and at least 100 must stay unreserved.
A new large reservation on another function, or another function scaling hard on the unreserved pool, can leave functions without a reservation nothing to scale into.
Detection
Metrics, logs and confirmation commands.
Root cause
Why the incident occurred.
Contain and prevent
Blast radius
What else this can affect.
Prevention
Controls and guardrails that stop it recurring.
-
Alarm on the Throttles metric for every function that serves customers, since Errors will not show this failure.
-
Keep reserved concurrency values in infrastructure code with a comment on why each exists, so the next person knows whether a 0 or a low ceiling is deliberate.
-
Review PutFunctionConcurrency20171031 and DeleteFunctionConcurrency20171031 events in CloudTrail after each release, and keep a trail if you need more than 90 days.
-
Track ClaimedAccountConcurrency against the account limit, and size new reservations with the requests-per-second times duration estimate before they go in.
Fix
Pre-flight checks
Check these before you change anything.
-
AWS CLI access in the function's account and Region.
-
Read access to Lambda, CloudWatch and CloudTrail; lambda:PutFunctionConcurrency or lambda:DeleteFunctionConcurrency for the change.
-
Agreement from the function owner before changing reserved concurrency.
Commands
aws lambda get-function-concurrency --function-name my-function
aws lambda get-account-settings
aws cloudwatch get-metric-statistics --namespace AWS/Lambda --metric-name Throttles --dimensions Name=FunctionName,Value=my-function --start-time <start-utc> --end-time <end-utc> --period 60 --statistics Sum
aws cloudtrail lookup-events --lookup-attributes AttributeKey=EventName,AttributeValue=PutFunctionConcurrency20171031 --start-time <start-utc> --max-items 50
aws lambda put-function-concurrency --function-name my-function --reserved-concurrent-executions <previous-or-sized-value>
aws lambda delete-function-concurrency --function-name my-function
Verify and recover
Verification
Evidence required before you close the incident.
-
aws lambda get-function-concurrency --function-name my-function — ReservedConcurrentExecutions equals the value set — The setting matches the change made.
-
Re-run the Throttles get-metric-statistics query for the minutes after the change — Sum of 0 in each 1-minute period — No throttles for at least 15 minutes of normal traffic.
-
Re-run the ConcurrentExecutions query with Maximum — Peak above the old ceiling or below the new reservation — The function is no longer pinned at its limit.
Commands in these steps
aws lambda get-function-concurrency --function-name my-function — ReservedConcurrentExecutions equals the value set — The setting matches the change made
Rollback
The safe way back.
-
Run aws lambda put-function-concurrency --function-name my-function --reserved-concurrent-executions <saved-value> with the value recorded by get-function-concurrency before the change.
-
If no reservation existed before the change, run aws lambda delete-function-concurrency --function-name my-function.
-
Confirm with aws lambda get-function-concurrency --function-name my-function that the saved state is back.
Escalate when
Conditions that need another owner.
-
ClaimedAccountConcurrency is at the account limit: request a higher concurrent executions quota through Service Quotas, and tell the owners of other functions in the account.
-
A 0 reservation was set to stop a recursive loop or protect a dependency: the function owner decides when it is safe to lift.
-
Throttles continue with the reservation restored and the account pool free, or put-function-concurrency keeps failing: open an AWS Support case with the saved outputs and timestamps.
-
The change came from an identity nobody can account for: hand over to security before restoring anything else. flowchart TD A[Lambda function returns TooManyRequestsException] --> B[Capture reserved concurrency, account settings and Throttles] B --> C{Reserved concurrency set on the function?} C -- Set to 0 --> D{Owner set 0 to stop a loop or protect a dependency?} D -- Yes --> E[Escalate and keep the throttle until the owner clears it] D -- No --> F[Put back the previous reserved concurrency value] C -- Set above 0 --> G{ConcurrentExecutions peaks at the reserved value?} G -- Yes --> H[Raise reserved concurrency, leaving 100 unreserved] G -- No --> I{ClaimedAccountConcurrency at the account limit?} C -- Not set --> I I -- Yes --> J[Escalate for a higher account concurrency quota] I -- No --> K[Open an AWS Support case with the saved outputs] F --> V[Verify Throttles falls to 0] H --> V Account owner.
Fixed? Get the next one before it hits you: The KBY Brief
Practical fixes for the systems you run, by email about twice a week. Unsubscribe at any time.
About this page
Credited to
Dates
Published
Corrections
No corrections are listed for this page. Some review sweeps corrected pages without naming each one; those are in the corrections log.
Was this useful?
Reports go to the site owner and are used to correct pages. Corrections are listed at /corrections.
Sources
Based on the vendor sources listed below. Check each step against your environment before acting.
First published 17 November 2026
Configuring reserved concurrency for an Event Function
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
Invoke - AWS Lambda API Reference
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
PutFunctionConcurrency - AWS Lambda API Reference
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
GetFunctionConcurrency - AWS Lambda API Reference
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
DeleteFunctionConcurrency - AWS Lambda API Reference
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
GetAccountSettings - AWS Lambda API Reference
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
get-function-concurrency - AWS CLI Command Reference
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
delete-function-concurrency - AWS CLI Command Reference
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
put-function-concurrency - AWS CLI Command Reference
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
get-account-settings - AWS CLI Command Reference
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
Monitoring concurrency - AWS Lambda
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
Types of metrics for Lambda functions
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
Viewing CloudWatch metrics for Lambda functions
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
get-metric-statistics - AWS CLI Command Reference
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
Troubleshoot invocation issues in Lambda
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
Error handling and automatic retries for asynchronous invocation
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
Lambda quotas
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
Logging AWS Lambda API calls using AWS CloudTrail
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
Looking up CloudTrail events with the AWS CLI
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.
Actions, resources, and condition keys for AWS Lambda
Cited for: Supports the documented platform behaviour, diagnosis or recovery controls used by this Runbook.