Lambda Managed Instances: The Logs You Need
LMI-Capacity-Providers write lifecycle logs to CloudWatch. Structured JSON, default off-the blind spot between Function and Fleet gets smaller.
Lambda Managed Instances now receive Capacity Provider logs in CloudWatch – structured as JSON by default. Teams running Serverless on EC2 can finally track launches, terminations, and health checks without having to build their own observability project.
The most important points at a glance
- Logs enabled by default. Capacity Providers write lifecycle events as structured JSON logs to CloudWatch.
- Visibility into the EC2 layer. Launches, terminations, and health checks become filterable – provisioning errors can be isolated more quickly.
- For predictable workloads. LMI targets high, predictable workloads with specialized compute profiles and EC2 pricing levers such as Savings Plans.
Related:CloudFront 5xx: What VPC Origins teams need to check
Lambda Managed Instances (LMI) promise the best of both worlds: the Lambda operational model and EC2 control. In practice, the missing layer has often been exactly what’s happening in the background when a function scales in unexpected ways – and that’s precisely what the Capacity Provider addresses.
Why LMI was risky without logs
LMI (Lambda Managed Instances) runs Lambda functions on EC2 instances provisioned by Capacity Providers. This is attractive for high, predictable workloads and specialized compute. The blind spot: when provisioning fails, you see function errors, but not the host lifecycle.
That’s exactly where the new Capacity Provider logs come in. AWS writes them automatically to CloudWatch Logs. Structured JSON enables filtering and alerts without having to build custom log parsers.
What the logs specifically cover
The focus is on scaling activity and instance lifecycle: launches, terminations, health checks. Teams can identify failed operations and provisioning errors more quickly through these logs. They are available in all commercial AWS Regions where LMI is running. For all Capacity Providers, they are enabled by default.
Configuration remains possible via the Lambda Console, API, CLI, SAM, or CloudFormation. Standard CloudWatch Logs costs apply. This is a standard telemetry path that must be budgeted for-similar to Function Logs.
Five checks before LMI scales into production
- Inventory Capacity Providers. Which providers, which instance families, and what limits apply?
- Configure log groups and retention policies. “Default-on” does not mean “store forever.” Define retention periods and cost budgets.
- Set alarms for lifecycle failures. Treat failed launches and repeated terminations as pager events, not as user tickets.
- Correlate with function metrics. Analyze throttles, duration, and errors in conjunction with provider logs.
- Review pricing levers. Savings Plans and Reserved Instances only make sense once the instance layer is stable and observable.
Who benefits from this move
LMI (Lambda Managed Instances) isn’t a replacement for classic Lambda when dealing with spiky, infrequent workloads. Instead, it’s a tool designed for predictable, high-volume compute tasks with specialized requirements. The new logs make the operational reality more transparent: you get the serverless experience, but with host-level visibility in the logs.
If you’re already running LMI and hunting for incidents “somewhere between the Function and the Fleet,” you should make the Capacity Provider view in the Lambda Console-and its corresponding CloudWatch filters-a standard part of your runbook starting today.
Source: AWS What’s New – Lambda Managed Instances capacity provider logs (July 24, 2026).
Frequently Asked Questions
What are Lambda Managed Instances?
LMI runs Lambda functions on EC2 instances managed by Capacity Providers. The goal is to support high, predictable workloads with specialized compute profiles and EC2 pricing options-while maintaining Lambda-like operational simplicity.
What’s new about Capacity Provider logs?
AWS now publishes structured JSON logs for scaling events and instance lifecycle activities (launches, terminations, health checks) to CloudWatch. These logs are enabled by default and available in all commercial AWS Regions where LMI is offered.
Are there extra charges?
Standard Amazon CloudWatch Logs charges apply for telemetry. The logs themselves are a built-in feature of Capacity Providers and not a separate add-on product.
How do I get started with monitoring?
Open the Capacity Provider page in the Lambda Console, review the log groups, set retention policies, and create alarms for failed lifecycle events. Then correlate function metrics and provider logs within the same incident board.
Editor’s Picks
cloudmagazinCloudFront 5xx Errors: What VPC Origins Teams Need to CheckcloudmagazinNGINX Vulnerability: Ingress and Gateway Under Patch PressurecloudmagazinEight Minutes to AWS Admin Access: Open S3 Bucket LeakMore from the MBF Media Network
MyBusinessFutureWhen the Credit Workflow Stalls at the InboxDigital ChiefsOrphaned Access: The Silent Cybersecurity GapSecurityTodayServiceNow RCE: Six Checks Before Opening a TicketImage source: AI-generated (July 2026)
