Per-Region Lambda limits | Overview.

Per-Region Lambda limits - Overview.

Scope:

• Intro,
• Per-Region Lambda Limits for Compute & Storage Limits (Soft; can be increased),
• Per-Region Lambda Limits for Deployment & Payload Constraints (Hard; non-adjustable),
• Scaling & Invocation Limits,
• Function Count,
• Summary Table,
• Final Thoughts.

Intro:

• Here's twtech overview of Per-Region Lambda limits.
• Information for official AWS documentation.

https://docs.aws.amazon.com/lambda/latest/dg/gettingstarted-limits.html?utm_source=chatgpt.com

Per-Region Lambda Limits For:

1. Compute & Storage Limits (Soft; can be increased)

• Concurrent executions: Default quota is 1,000 concurrent executions per region, 
•  Concurrent executions is shared across all functions. 
• However, twtech can request an increase.
• Function & Layer storage: Up to 75 GB of storage per region.
• Function & Layer storage is for twtech deployment packages and layer versions, adjustable upon request.
• Elastic Network Interfaces (ENIs) per VPC: Default is 500 ENIs.
• Elastic Network Interfaces (ENIs) per VPC is also adjustable.

2. Deployment & Payload Constraints (Hard; non-adjustable)

• Maximum deployment package size:
• 50 MB zipped for direct uploads.
• 250 MB unzipped content (via S3 upload).
• Container image support: Supports container images up to 10 GB (uncompressed) stored in Amazon ECR.
• Layers per function: Up to 5 layers can be attached to a function.
• Temporary storage (/tmp): Limited to 512 MB ephemeral disk space per function.
• Environment variables: Combined size limit of 4 KB.
• Function timeout: Maximum timeout is 900 seconds (15 minutesProcesses/threads & file descriptors): 
• Function timeout supports Up to 1,024 threads/processes, and 1,024 open file descriptors per function.
• API request rates (control plane):
• GetFunction: 100 requests/sec.
• GetPolicy: 15 requests/sec.
• Other control-plane calls: 15 requests/sec.
• Payload size limits:
• Synchronous (request or response): Max 6 MB.
• Asynchronous payloads: Max 256 KB. 

3. Scaling & Invocation Limits

• Concurrent scaling rate: Per function, scales at up to 1,000 additional execution environments every 10 seconds (or 10,000 additional requests/sec); 
• Concurrent scaling rate cannot accumulate unused capacity.
• Overall requests/sec cap: Lambda additionally enforces a limit of 10× the account concurrency (i.e., 10,000 requests/sec by default with 1,000 concurrency). 
• Burst throttling: Sudden cold-start spikes can be subject to a burst limit via token-bucket algorithms, limiting how quickly concurrency can ramp up even before hitting maximum concurrency.

4. Function Count

• There is no fixed limit on the number of Lambda functions twtech can create per region. 
• However, the 75 GB storage limit for function & layer code acts as an implicit cap based on average package size.

Summary Table

Limit Type	Per Region Default	Adjustable?
Concurrent executions.	1,000	Yes
Function & layer storage.	75 GB	Yes
ENIs per VPC.	500	Yes
Deployment package (zipped).	50 MB	No
Deployment package (unzipped).	250 MB	No
Container image size.	10 GB	No
Layers per function.	5	No
/tmp storage.	512 MB	No
Env vars size.	4 KB	No
Timeout.	15 minutes	No
Threads/File descriptors.	1,024	No
Payload size (sync).	6 MB	No
Payload size (async).	256 KB	No
Scaling rate per function.	1,000 envs / 10s	No
Requests/sec limit.	10× concurrency	No
Burst/concurrency ramp-up.	Token-bucket throttling	No
Number of functions.	Unlimited (storage bound)	—

twtech Final Thoughts

• These quotas are enforced per AWS Region.
•  Each region has its own limits. 
• Some limits are soft (increaseable)—like concurrency and storage.
• while other limits are hard caps, meaning twtech must design around e.g., payload size, timeout, temp storage.