Glue Job Cost Calculator
Price AWS Glue ETL jobs per DPU-hour with the 1-minute minimum billing, worker types, retries and monthly run count.
Inputs
Glue ETL is around $0.44 per DPU-hour in us-east-1, billed per second.
Glue 2.0 and later bill a 1-minute minimum per run; Glue 1.0 billed 10 minutes.
Estimated Monthly Cost
$116.16
DPUs Allocated per Run
20DPU
Billed Duration per Run
12.00minutes
DPU-hours per Month
264.00DPU-hours
Cost per Run
$1.76
Estimated Annual Cost
$1,393.92
Step by step
Values used
Workers = 10 workers; Worker type = G.2X — 2 DPU, 8 vCPU, 32 GB; Job run time = 12 minutes; Runs per month = 60 runs; Runs that fail and retry = 10 %; DPU price = 0.4400 $/DPU-hour; Minimum billed duration = 1 minutes
Glue Job Cost
DPU-hours per run = workers × DPU per worker × max(run minutes, minimum minutes) ÷ 60; monthly = DPU-hours × price × runs × (1 + retry rate).
Estimated Monthly Cost
= 116.16
DPUs Allocated per Run
= 20 DPU
Billed Duration per Run
= 12.00 minutes
DPU-hours per Month
= 264.00 DPU-hours
Cost per Run
= 1.76
Estimated Annual Cost
= 1,393.92
How it works
Glue bills allocated DPU-hours per second with a one-minute floor, so cost is driven by workers times DPU-per-worker times wall-clock time — doubling the workers halves the runtime and costs the same only if the job scales perfectly. Failed runs are billed for the time they consumed before failing, which is why the retry rate belongs in the estimate. The DPU price is an editable input with a realistic us-east-1 default, so confirm the current Glue rate with AWS for your region and job type. The one-minute minimum makes many small Glue jobs far more expensive per byte than one larger job, and G.2X workers with fewer, longer tasks usually beat G.1X workers with more shuffle.
Formula
Glue Job Cost
DPU-hours per run = workers × DPU per worker × max(run minutes, minimum minutes) ÷ 60; monthly = DPU-hours × price × runs × (1 + retry rate).
- DPU
- Data Processing Unit — 4 vCPU and 16 GB of memory
- minimum billed duration
- Floor of 1 minute per run on Glue 2.0 and later
- retry rate
- Share of runs that fail and are re-executed, and billed again
Frequently Asked Questions
How is Glue Job Cost calculated?
DPU-hours per run = workers × DPU per worker × max(run minutes, minimum minutes) ÷ 60; monthly = DPU-hours × price × runs × (1 + retry rate). Glue bills allocated DPU-hours per second with a one-minute floor, so cost is driven by workers times DPU-per-worker times wall-clock time — doubling the workers halves the runtime and costs the same only if the job scales perfectly. Failed runs are billed for the time they consumed before failing, which is why the retry rate belongs in the estimate. The DPU price is an editable input with a realistic us-east-1 default, so confirm the current Glue rate with AWS for your region and job type.
Why does Glue Job Cost matter?
The one-minute minimum makes many small Glue jobs far more expensive per byte than one larger job, and G.2X workers with fewer, longer tasks usually beat G.1X workers with more shuffle.
What values do I need to enter?
This calculator takes 7 inputs: Workers, Worker type, Job run time, Runs per month, Runs that fail and retry, DPU price, Minimum billed duration. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.
Should I use more workers to save money?
Only if the job scales. Cost is DPU-hours, so doubling workers is free when runtime halves — but shuffle-heavy or skewed jobs stop scaling and you pay double for the same wall-clock time. Check the Spark UI for straggler tasks before adding workers.
Are these prices current?
The $0.44 per DPU-hour default is a realistic us-east-1 figure for Glue ETL, but rates differ by region and by job type such as Python shell, streaming or Flex. Confirm on the AWS Glue pricing page.