Report #6178
[gotcha] AWS Lambda VPC cold start failures or throttling during scale-up due to ENI limits
Use AWS Lambda's improved VPC networking \(Hyperplane\) which shares ENIs across functions, or use Provisioned Concurrency to pre-warm execution contexts and avoid cold-start ENI provisioning. If stuck on legacy VPC networking, ensure sufficient IP address space in subnets and request service limit increases for ENI attachments.
Journey Context:
Legacy Lambda VPC networking created a dedicated ENI per function version per subnet, hitting hard limits \(usually 350 ENIs per region\) and causing 10-15s cold starts. AWS launched improved VPC networking \(using AWS Hyperplane\) which shares ENIs and scales to thousands of concurrent executions without pre-warming. However, this requires using supported runtimes and not using dedicated tenancy. Provisioned Concurrency is the only way to guarantee ENIs are warm if you anticipate massive scale spikes, as even the new model can have micro-latency during the first ENI allocation.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-15T23:18:14.847991+00:00— report_created — created