Lambda Storage Shift: S3 Takes the Wheel, Not the Limit

Lambda's Storage Revolution

AWS Lambda's introduction of self-managed code storage in S3 is a welcome evolution, primarily addressing the long-standing account-level code storage quota that often required manual intervention via support tickets. By allowing developers to leverage their own S3 buckets, this feature effectively removes that aggregate ceiling, offering a more scalable solution for teams managing extensive function fleets. The increased default Lambda-managed storage to 300 GB is also a positive step, though the core benefit lies in the move to customer-owned storage. This shift provides greater transparency and control, enabling the application of existing S3 metrics and lifecycle policies to Lambda code, which can be a significant operational advantage for cost management and organization. Furthermore, the elimination of an intermediate code copy by Lambda can lead to faster function activation times, a subtle but impactful performance improvement.

However, it's crucial to understand the nuanced implications. The announcement's framing, as highlighted in the article, can be misleading. While the account storage quota is effectively lifted, the per-function size limits remain unchanged. This means that developers still cannot circumvent the 50MB zipped or 250MB unzipped limits for zip-based functions, nor the 10GB limit for container images, by simply moving their code to S3. The innovation here is in the management and allocation of storage, not in increasing the fundamental size constraints of individual Lambda functions. The operational change involves a shift in billing from an implicit Lambda cost to explicit S3 storage and retrieval charges, which, while more transparent, requires careful consideration. The lack of immediate Infrastructure-as-Code (IaC) support, particularly for Terraform, is a notable drawback for teams heavily invested in automated deployments, necessitating workarounds until provider updates are available. This feature is particularly beneficial for large enterprises, agencies, and any organization with a substantial number of Lambda functions that were previously bumping against or anticipating hitting the account storage quota.

Key Points

• AWS Lambda now supports self-managed code storage, allowing deployment packages to be referenced directly from customer-owned S3 buckets.
• This change removes the per-Region code storage quota and raises the default Lambda-managed storage to 300 GB.
• Individual function size limits (50MB zipped, 250MB unzipped for zip; 10GB for container images) remain unchanged.
• The shift offers greater operational control, transparency, and the ability to use S3 features like metrics and lifecycle rules for Lambda code.
• Lambda no longer creates an intermediate copy of the package, potentially speeding up function activation after updates.
• Costs shift from an invisible Lambda quota to visible S3 storage and retrieval charges.
• Infrastructure-as-Code support (e.g., Terraform) is still evolving for this feature.

---

📖 Source: AWS Lambda's Self-Managed Code Storage Lifts the Account Quota, Not the Function Size Limit