Skip to main content

Lambda update function memory

Last updated on

Lambda update function memory is an AWS chaos fault that changes the memory allocation of a target Lambda function (FUNCTION_NAME in REGION) to MEMORY_IN_MEGABYTES for TOTAL_CHAOS_DURATION seconds, then restores the original value. Because Lambda allocates CPU proportionally to memory, lowering memory also reduces the CPU share available to each invocation.

Use this fault to test how a Lambda workload behaves under a tightened memory and CPU budget: whether invocations OOM, whether duration rises, whether downstream consumers degrade gracefully, and whether monitoring detects the resource pressure before it cascades.

Run your first experiment

If you have not configured the chaos infrastructure yet, go to Quickstart to install the chaos infrastructure and run an experiment end to end.

Use cases​

Run this fault when you want to answer concrete questions like:

  • OOM behaviour: When memory drops to MEMORY_IN_MEGABYTES, do invocations fail with Runtime exited with error: signal: killed or proceed slowly?
  • Duration impact: How much does the reduced CPU share extend Duration, and does it push the function past its configured timeout?
  • Caller impact: Do synchronous callers (API Gateway, ALB) see timeouts and elevated 5xx?
  • Monitoring fidelity: Do CloudWatch alarms on Errors, Duration, and downstream p99 fire within the SLA?
  • Capacity planning: Identify the lowest safe memory setting for cost optimization without breaking the SLA.

Prerequisites​

  • Kubernetes version: 1.21 or later for the chaos infrastructure cluster. Go to What's supported to confirm distribution support.
  • Target Lambda function: FUNCTION_NAME exists in REGION and is in the Active state.
  • Memory value: MEMORY_IN_MEGABYTES is within Lambda's allowed range (128 to 10240, in 1 MB increments).
  • AWS credentials available: Either an AWS credentials file uploaded as a File Secret in Harness Secret Manager (see Authentication below) or an IAM role for service accounts (IRSA) bound to the chaos infrastructure service account.
  • IAM permissions granted: The credentials or role include the permissions listed below.

Supported environments​

PlatformSupport status
AWS Lambda (zip and container image)Supported
Lambda with provisioned concurrencySupported (provisioned concurrency is rebuilt at the new memory)
Lambda@EdgeNot supported
AWS regionsSupported in every commercial region; pass the region in REGION

Permissions required​

The IAM principal that the chaos pod uses (the credentials mounted from the Harness Secret Manager file secret, the IRSA role on the chaos service account, or the role assumed via ASSUME_ROLE_ARN) needs the following AWS actions.

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "lambda:GetFunction",
        "lambda:GetFunctionConfiguration",
        "lambda:UpdateFunctionConfiguration"
      ],
      "Resource": "*"
    }
  ]
}

Go to common policy for all AWS faults to use a single superset IAM policy across every AWS fault.

Authentication​

The fault supports three credential delivery models. Pick one based on how your chaos infrastructure is deployed.

MethodWhen to use itHow to configure
Harness Secret Manager file secretChaos infrastructure runs outside EKS, or you want explicit static credentialsUpload the AWS credentials file as a File Secret in Harness Secret Manager and reference its identifier via AWS_AUTHENTICATION_SECRET
IAM Roles for Service Accounts (IRSA)Chaos infrastructure runs in EKS and uses an OIDC-bound service accountNo tunable changes; the chaos pod inherits the role automatically. Go to AWS IAM integration to set it up
Assume roleThe fault needs to act in a different account or with elevated permissionsSet ASSUME_ROLE_ARN to the role ARN; the chaos pod assumes the role on top of its base credentials

When using the Harness Secret Manager method, the contents of the File Secret should be the AWS credentials file in the standard ~/.aws/credentials format:

[default]
aws_access_key_id = REPLACE_WITH_ACCESS_KEY_ID
aws_secret_access_key = REPLACE_WITH_SECRET_ACCESS_KEY

Upload this file as a File Secret in Harness Secret Manager (Project Setup → Secrets → New File Secret), and pass the secret identifier in AWS_AUTHENTICATION_SECRET when configuring the fault.

Fault tunables​

Configure the following fault parameters when you add Lambda update function memory to an experiment in Chaos Studio. Defaults are shown for reference.

Required parameters

TunableDescriptionDefault
FUNCTION_NAMEName of the target Lambda function.(required)
REGIONAWS region where the Lambda function is deployed.(required)
MEMORY_IN_MEGABYTESMemory size (in MB) to set on the function during the chaos window. Allowed range 128 to 10240.(required)

Chaos parameters

TunableDescriptionDefault
TOTAL_CHAOS_DURATIONDuration of the fault in seconds.30
CHAOS_INTERVALDelay in seconds between successive iterations when running for more than one cycle.30
RAMP_TIMEWait period in seconds before and after the fault. Go to ramp time to read how it is applied.0

Authentication

TunableDescriptionDefault
ASSUME_ROLE_ARNARN of an IAM role to assume on top of the base credentials. Leave empty to use the base credentials directly.""
AWS_AUTHENTICATION_SECRETIdentifier of the File Secret in Harness Secret Manager that contains the AWS credentials file. Not required when using IRSA.""

Tunables that apply to every fault are documented in common tunables for all faults.

Fault execution in brief​

Reads the current memory allocation on FUNCTION_NAME, updates it to MEMORY_IN_MEGABYTES, waits for TOTAL_CHAOS_DURATION seconds, then restores the original memory value.

Expected behavior during fault execution​

  • New invocations of the function run with MEMORY_IN_MEGABYTES of RAM and a proportionally smaller CPU share.
  • Workloads that previously fit comfortably may OOM (Runtime exited with error: signal: killed) or run noticeably slower.
  • CloudWatch Errors rises for OOM-killed invocations; Duration rises for compute-bound invocations.
  • Warm containers initialized before the update continue to run with the original memory until they are recycled.
When the fault ends

The chaos pod restores the original memory value. New invocations run with the previous allocation. Warm containers carrying the reduced memory are replaced on the next cold start.

Signals to watch​

Attach resilience probes to assert each layer:

  • Lambda duration: Use a Prometheus probe on the duration metric for the function.
  • Lambda errors: Use a Prometheus probe on aws_lambda_errors_sum.
  • End-to-end availability: Use an HTTP probe on the user-visible endpoint.

Verify the fault execution effect​

While the experiment is running, confirm the memory was reduced and then restored:

  1. Inspect function configuration.

    aws lambda get-function-configuration \
      --region <region> \
      --function-name <name> \
      --query MemorySize

    The reported value should equal MEMORY_IN_MEGABYTES during the chaos window and return to the original value afterwards.

  2. Invoke the function and inspect the log report.

    aws lambda invoke --region <region> --function-name <name> /tmp/out.json
    aws logs tail --region <region> /aws/lambda/<name> --since 5m

    REPORT lines should show Memory Size: <MEMORY_IN_MEGABYTES> MB and may show Max Memory Used close to the new limit; OOM cases appear as Runtime exited with error: signal: killed.

Recovery and cleanup​

  • End of duration: The chaos pod restores the original memory value.
  • Abort the experiment: Stopping the experiment from Chaos Studio also triggers the restore call.
  • Manual recovery: If the fault exits before restore, run aws lambda update-function-configuration --function-name <name> --memory-size <N> --region <region> with the original value (recorded in the chaos pod logs).
  • Workload recovery: Warm containers carrying the reduced memory are replaced on the next cold start.

Limitations​

  • Memory and CPU bundled: Lambda allocates CPU proportionally to memory. Reducing memory also reduces CPU; you cannot isolate one from the other.
  • Provisioned concurrency rebuilds: Updating memory rebuilds provisioned concurrency at the new size, which incurs init time.
  • Allowed range: MEMORY_IN_MEGABYTES must be between 128 and 10240. Values outside this range are rejected by AWS.
  • Cold-start interaction: Updating the function configuration triggers cold starts during the chaos window.
  • Cross-region targeting: A single experiment targets one region (the value of REGION).

Troubleshooting​

Lambda update function memory fails with AccessDeniedException in Harness Chaos Engineering

The credentials supplied to the chaos pod do not have the required Lambda permissions. Confirm the IAM policy attached to the user, role, or IRSA service account includes lambda:GetFunction, lambda:GetFunctionConfiguration, and lambda:UpdateFunctionConfiguration.

MEMORY_IN_MEGABYTES value rejected

Lambda allows memory between 128 and 10240 MB in 1 MB increments. Re-run the experiment with a value inside this range.

No measurable impact at the reduced memory

If the workload was already running comfortably below MEMORY_IN_MEGABYTES, lowering the limit may have no visible effect. Drive sustained load against the function and pick a memory value below the typical Max Memory Used reported in CloudWatch Logs.

Memory not restored after the chaos window

If the restore call failed, the function may stay at the reduced memory. Run 'aws lambda update-function-configuration --function-name <name> --memory-size <N> --region <region>' with the original value recorded in the chaos pod logs.