custom-scaling

Custom scaling example

Same VPC + security group layout as examples/basic, with python3.14 and the split modules:

lambda_managed_instance (fleet / capacity provider):

• scaling_mode = "Manual" with a CPU target (cpu_target_utilization) so the pool tracks average CPU utilisation
• allowed_instance_types — a pinned allowlist of x86_64 sizes (for example m7i.2xlarge, c7i.4xlarge) so placement stays predictable

lambda_managed_function (one function on that provider):

• reserved_concurrent_executions, per_execution_environment_max_concurrency, environment_variables, and structured logging (log_format, application_log_level)

Defaults use lmi-custom and 10.1.0.0/16 so you can run alongside basic (10.0.0.0/16) without overlapping CIDRs.

Apply

cd examples/custom-scaling
terraform init
terraform plan
terraform apply

Optional: add a terraform.tfvars file to override variables from variables.tf (defaults are set for a non-overlapping CIDR with examples/basic).

Tight allowed lists can reduce placement flexibility if a type is capacity-constrained in your region — keep at least two sizes in the allowlist where possible.

Invoke after apply

aws lambda invoke \
  --function-name "$(terraform output -raw lambda_function_name):$(terraform output -raw lambda_version)" \
  --payload '{"test":true}' \
  --cli-binary-format raw-in-base64-out /tmp/out.json && cat /tmp/out.json

Related examples

• examples/basic — defaults-only Auto scaling
• examples/waf-loki — walkthrough networking lab stack

Product reference: Scaling Lambda Managed Instances (instance type selection).