AWS Lambda: TooManyRequestsException: Rate Exceeded.
Lambda refused the invocation because the function hit its concurrency ceiling, which is the account's regional limit (1,000 by default, lower on new accounts) or the function's own reserved concurrency. It is a limit on requests running at the same moment, not requests per second, so a function that got slower can start throttling at the same traffic. The Reason field in the response says which limit it was.
Quick fix
Read the commands before running them. Anything that restarts a service, deletes data or changes permissions should be tried on a non-production system first.
# Is it the function's reserved limit or the account limit?
aws lambda get-function-concurrency --function-name my-fn
aws lambda get-account-settings --query 'AccountLimit.ConcurrentExecutions'
# Throttles and concurrency over the last hour
aws cloudwatch get-metric-statistics --namespace AWS/Lambda \
--metric-name Throttles --dimensions Name=FunctionName,Value=my-fn \
--start-time $(date -u -d '-1 hour' +%FT%TZ) --end-time $(date -u +%FT%TZ) \
--period 60 --statistics Sum
# Reserved concurrency of 0 means the function is effectively disabled
aws lambda put-function-concurrency --function-name my-fn \
--reserved-concurrent-executions 100
# Account limit: request an increase in Service Quotas
# Also check duration: slower invocations hold concurrency longerHow to diagnose Serverless errors
Serverless errors are shaped by hard platform limits: execution duration, deployment package size, memory, and concurrency. Unlike a server you control, these cannot be tuned away. The fix is to restructure the workload. VPC-attached functions add a second class of problem, because they lose default internet access and depend on ENI capacity and NAT configuration.
If the quick fix above does not resolve it, work through these steps. They apply to this whole class of error, not just to this one message, which is usually what saves the time.
- Check the platform limit before changing code: timeout, package size and memory are all documented, and the error is usually the limit being reported literally.
- For timeouts, add timing logs around each external call. A function that times out is nearly always waiting on one downstream dependency.
- Reduce package size with layers, tree-shaking, and by excluding dev dependencies and the provider SDK that the runtime already includes.
- For VPC functions with no internet access, confirm there is a NAT Gateway or the relevant VPC endpoints. A function in a private subnet cannot reach the public internet by default.
- Measure cold start separately from warm invocation in your metrics, and use provisioned concurrency only after confirming cold starts are the actual problem.
Tools worth reaching for
CloudWatch Logs Insightsaws lambda get-function-configurationX-Ray / distributed tracingsam local invoke
Authoritative references
Primary documentation for this error, worth reading before applying any fix in production.
Related Serverless errors
- API Gateway: {"message": "Endpoint request timed out"}API Gateway stopped waiting for the integration, which for REST APIs happens at 29 seconds by…
- AWS Lambda: ENI limit reached in VPCLambda functions in a VPC create ENIs (Elastic Network Interfaces) and the account or subnet…
- AWS Lambda: Internal server error, malformed Lambda proxy responseAPI Gateway received something it could not turn into an HTTP response. With proxy…
- AWS Lambda: Response payload size exceeded maximum allowed payload sizeSynchronous invocations have a fixed 6 MB limit on the response (and on the request), and it…
- AWS Lambda: Runtime exited with error: signal: killedThe function process was killed for exceeding its memory allocation. Lambda reports this…
- AWS Lambda: Runtime.HandlerNotFound: index.handler is undefined or not exportedLambda found and loaded the file named in the handler setting, but that file has no export…
- AWS Lambda: Runtime.ImportModuleErrorThe runtime could not load the handler module. Either the handler string does not match the…
- AWS Lambda: Runtime.InvalidEntrypoint (exec format error)Lambda could not start the container image's entrypoint at all. When the detail says exec…
Browse other categories
- HTTP 494xx client errors, 5xx server errors, redirects, headers and protocol problems.
- JavaScript 42npm resolution, async pitfalls, hydration, memory limits and runtime type…
- Database 41Connections, deadlocks, constraints, replication and memory limits.
- AI 35Rate limits, context windows, GPU memory and model-serving failures.
- Network 35Refused connections, timeouts, resets, MTU problems and port exhaustion.
- Python 35Imports, virtual environments, encoding, concurrency and dependency conflicts.
- Kubernetes 34CrashLoopBackOff, ImagePullBackOff, OOMKilled, RBAC, scheduling and storage.
- Docker 27Daemon connectivity, disk space, image pulls, ports and architecture mismatches.
- System 26Disk space, systemd units, file descriptors, OOM killer and scheduled jobs.
- Cloud 25IAM permissions, quotas, service limits and credential failures.
- Security 25JWT validation, CSRF, OAuth grants, SELinux, SSH host keys and CSP.
- TLS 24Untrusted authorities, expiry, hostname mismatch, chains and cipher negotiation.
Something missing or wrong?
This entry is maintained by hand. If the fix is out of date, incomplete, or you have a better one, email a correction and it will be reviewed.