Unexpected MaxCompute charges often stem from a single high-cost SQL job or a storage spike. This page explains how to find the source of a charge — from your billing summary down to the specific job — and what to look for along the way.
The investigation follows three steps:
-
View billing details — Identify which project and billable item drove the charge.
-
Analyze usage records — Download granular data to confirm the abnormal entry.
-
Query job history — Use Information Schema to find the specific SQL job and investigate the root cause.
View billing details
Open Expenses and Costs in the Alibaba Cloud Management Console and go to Billing Management to review your charges.
For pay-as-you-go MaxCompute projects, bills are generated before 07:00 the next day.
For step-by-step instructions, see View billing details.
Analyze usage records
If your daily bill spikes to a multiple of the normal amount, download usage records to identify the source.
To investigate a billing spike:
-
In Expenses and Costs, locate the project and billable items responsible for the increase.
-
Download the usage records for that project.
-
Review the records to identify the abnormal entries.
Each usage record contains the complexity and billing duration of an SQL statement, storage resource consumption, and download traffic in the last 24 hours.
For instructions, see Analyze the usage records of MaxCompute bills.
How storage fees are calculated
Keep the following in mind when reviewing storage charges:
Metering data is pushed hourly and collected 24 times per day.
The daily storage fee equals the average byte count over 24 hours, priced using tiered rates.
A task's metering is attributed to the day it completes, not the day it starts. Tasks that finish after midnight count toward the completion day.
Data transfer charges
| Transfer type | Charged? |
|---|---|
| Upload data | No |
| Download over cloud product interconnection network | No |
| Download over the Internet | Yes |
Query job history with Information Schema
When a compute job causes an unexpected charge, use Information Schema to examine job execution details and find the root cause.
Query the TASKS_HISTORY view to list completed jobs and identify the most expensive ones.
By default, Information Schema retains data for the last 14 days only.