Measure invocations, memory impact, duration, and scaling behavior. Export results, compare scenarios, and inspect trends. Make smarter serverless decisions using simple reliable calculations today.
| Scenario | Monthly Requests | Avg Duration (ms) | Memory (MB) | Architecture | Estimated Cost | Suggested Provisioned Concurrency |
|---|---|---|---|---|---|---|
| Starter API | 1,200,000 | 160 | 512 | ARM | $0.040000 | 2 |
| Media Worker | 5,000,000 | 450 | 2,048 | X86 | $72.281935 | 29 |
| Analytics Stream | 9,000,000 | 280 | 1,536 | ARM | $53.283970 | 30 |
| ML Inference | 3,000,000 | 900 | 4,096 | X86 | $175.750344 | 33 |
Effective Requests = Monthly Requests × Retry Multiplier
Duration in Seconds = Average Duration in Milliseconds ÷ 1000
Memory in GB = Memory in MB ÷ 1024
Total Compute Seconds = Effective Requests × Duration in Seconds
Total Compute GB-seconds = Total Compute Seconds × Memory in GB
Billable Requests = Effective Requests − Free Tier Requests
Billable Compute GB-seconds = Total Compute GB-seconds − Free Tier Compute
Request Cost = (Billable Requests ÷ 1,000,000) × Request Rate
Compute Cost = Billable Compute GB-seconds × Compute Rate
Billable Extra Ephemeral GB = (Ephemeral Storage MB − 512) ÷ 1024
Ephemeral GB-seconds = Billable Extra Ephemeral GB × Total Compute Seconds
Ephemeral Cost = Ephemeral GB-seconds × Ephemeral Rate
Average Concurrency = Average Requests Per Second × Duration in Seconds
Peak Concurrency = Peak Requests Per Second × Duration in Seconds
Suggested Provisioned Concurrency = Peak Concurrency × (1 + Buffer %)
Enter your expected monthly request count first. Add the average execution duration in milliseconds. Choose the memory size that matches your function configuration. Select x86 or Arm based on your deployment target.
Now enter the average requests per second and the peak requests per second. These values help estimate average concurrency and peak concurrency. Add a retry multiplier if your events are retried during failures or queue reprocessing.
Set ephemeral storage if your function uses temporary file space. Keep the free tier checkbox selected if you want a lighter estimate. Remove it when you need a raw cost view without that deduction.
Press the calculate button. The result appears above the form. Review request cost, compute cost, storage cost, concurrency values, and suggested provisioned concurrency. Use the export buttons to save a CSV file or a PDF copy.
AWS Lambda billing looks simple at first. The details change the final number. Request count matters. Execution time matters. Memory size matters even more. Temporary storage can also change cost. Concurrency planning affects stability and user experience. A small error in input can distort an entire monthly estimate. This calculator collects the main workload variables in one place. It turns raw usage assumptions into practical values. You can review cost, pressure, and scaling in a single view. That makes planning easier before deployment.
Serverless projects often grow fast. Early estimates are usually optimistic. Teams focus on code and forget request spikes. They also forget retries, event duplication, and burst traffic. This calculator solves that gap with direct workload math. It computes effective invocations after retries. It converts milliseconds into seconds. It converts memory into gigabytes. Then it measures total compute in GB-seconds. That value drives billing. Free tier deductions can be applied when needed. You can also test storage-heavy functions that use ephemeral disk space during processing.
Cost is only part of the story. Performance is another part. Slow functions increase concurrency demand. High traffic peaks amplify that problem. This calculator uses average requests per second and peak requests per second to estimate real concurrency. That is useful during launch planning. It also helps when traffic is uneven across the day. A buffer percentage adds practical headroom. The suggested provisioned concurrency value gives a safer target. That makes capacity planning clearer. It also helps compare memory changes against speed and monthly spend.
The best use of this tool is scenario testing. Start with your current estimates. Then test a faster duration. Test a larger memory size. Test Arm and x86 separately. Test more retries. Test higher peak traffic. These small comparisons show where cost comes from. They also show where risk appears. Export the result as CSV for documentation. Save the PDF when sharing with clients or teams. The graph helps explain the cost mix quickly. That turns vague planning into a structured deployment decision.
It estimates request charges, compute charges, ephemeral storage charges, average concurrency, peak concurrency, and suggested provisioned concurrency from your workload inputs.
Lambda pricing depends on execution time and allocated memory. More memory increases available compute power, but it can also raise the billed GB-second total.
GB-seconds measure memory allocation multiplied by execution time. It is a core billing unit for Lambda compute charges.
Retries increase the real number of executions. A retry multiplier helps model event reprocessing, transient failures, and queue-driven invocation repeats.
Concurrency shows how many requests run at the same time. It helps estimate scale pressure and supports safer planning for bursts.
Yes. Keep the checkbox selected to deduct the free tier estimate. Clear it when you want a direct raw workload calculation.
Extra ephemeral storage above the included amount can add cost. This matters for workloads that create temporary files, media output, or model data.
Yes. Change the architecture field and recalculate. This helps compare estimated monthly cost and scaling outcomes with the same traffic pattern.
Important Note: All the Calculators listed in this site are for educational purpose only and we do not guarentee the accuracy of results. Please do consult with other sources as well.