Serverless Cost Calculator

Compare AWS Lambda, Cloudflare Workers, and Vercel Functions costs across different traffic patterns. Includes cost formulas, Lambda Power Tuning, and cold start optimization strategies.

Platform Cost Comparison

Lambda Cost Formula

# Lambda cost calculation
# Cost = (requests × $0.0000002) + (GB-seconds × $0.0000000166667)
# GB-seconds = (memory_mb / 1024) × duration_ms / 1000 × monthly_requests

def calculate_lambda_cost(monthly_requests, memory_mb, avg_duration_ms):
    request_cost = monthly_requests * 0.0000002
    gb_seconds = (memory_mb / 1024) * (avg_duration_ms / 1000) * monthly_requests
    compute_cost = gb_seconds * 0.0000000166667

    # Subtract free tier
    free_requests = 1_000_000
    free_gb_seconds = 400_000
    billable_requests = max(0, monthly_requests - free_requests)
    billable_gb_seconds = max(0, gb_seconds - free_gb_seconds)

    total = (billable_requests * 0.0000002) + (billable_gb_seconds * 0.0000000166667)
    return {
        "monthly_requests": monthly_requests,
        "gb_seconds": round(gb_seconds, 2),
        "monthly_cost_usd": round(total, 4)
    }

# Examples
configs = [
    (1_000_000, 128, 100),
    (10_000_000, 256, 200),
    (100_000_000, 512, 50),
]
for req, mem, dur in configs:
    result = calculate_lambda_cost(req, mem, dur)
    print(f"{req/1e6:.0f}M req @ {mem}MB/{dur}ms = ${result['monthly_cost_usd']}/month")

Lambda Power Tuning Setup

# Deploy Lambda Power Tuning via SAR
aws serverlessrepo create-cloud-formation-change-set \
  --application-id arn:aws:serverlessrepo:us-east-1:451282441545:applications/aws-lambda-power-tuning \
  --stack-name lambda-power-tuning \
  --capabilities CAPABILITY_IAM \
  --parameter-overrides '[{"name":"lambdaResource","value":"*"}]'

# After deploy, invoke the Step Functions state machine
aws stepfunctions start-execution \
  --state-machine-arn "arn:aws:states:us-east-1:YOUR_ACCOUNT:stateMachine:powerTuningStateMachine" \
  --input '{
    "lambdaARN": "arn:aws:lambda:us-east-1:YOUR_ACCOUNT:function:my-api-handler",
    "powerValues": [128, 256, 512, 1024, 2048, 3008],
    "num": 10,
    "payload": "{\"path\":\"/health\"}",
    "parallelInvocation": true,
    "strategy": "cost"
  }'

# strategy options: "cost" | "speed" | "balanced"
# Results visualized at: https://lambda-power-tuning.show/

Memory vs Duration Tradeoff

Higher memory often runs faster enough that total GB-seconds is similar or lower. Always measure with Power Tuning — never guess.

Cold Start Mitigation

# Option 1: Provisioned Concurrency (eliminates cold starts, costs more)
aws lambda put-provisioned-concurrency-config \
  --function-name my-api-handler \
  --qualifier prod \
  --provisioned-concurrent-executions 10

# Option 2: EventBridge keep-warm rule (free, reduces cold starts)
# Create rule to invoke every 5 minutes
aws events put-rule \
  --name KeepWarmRule \
  --schedule-expression "rate(5 minutes)"

aws lambda add-permission \
  --function-name my-api-handler \
  --statement-id KeepWarmPermission \
  --action lambda:InvokeFunction \
  --principal events.amazonaws.com \
  --source-arn arn:aws:events:us-east-1:YOUR_ACCOUNT:rule/KeepWarmRule

# Option 3: Lambda SnapStart (Java 21 runtime)
# In function configuration:
# SnapStart: ApplyOn: PublishedVersions
# Then invoke by version alias, not $LATEST