Data Lake Formation (DLF) charges for three billable items: storage of metadata objects, metadata requests, and resource usage for data ingestion.
All DLF features are currently free of charge. No charges apply even if your usage exceeds the free tiers described below.
Billable items
The following table summarizes each billable item, its free tier, and the rate above the free tier.
| Billable item | Free tier | Rate above free tier |
|---|---|---|
| Storage of metadata objects | First 1,000,000 objects/month | USD 1 per 100,000 objects/month |
| Metadata requests | First 1,000,000 requests/month | USD 1 per 1,000,000 requests/month |
| Data ingestion | Free (tasks have no charge) | N/A |
Storage of metadata objects
Metadata objects are the databases, tables, partitions, and indexes stored in your data lake's metadata. Every object in these four categories counts toward your monthly total.
Pricing
| Usage | Price |
|---|---|
| First 1,000,000 objects/month | Free |
| Above 1,000,000 objects | USD 1 per 100,000 objects/month |
-
Billing rule: The first 1 million stored objects each month are free. For usage that exceeds this free tier, you are charged USD 1 per month for every 100,000 additional objects. Any fraction of 100,000 objects is billed as a full 100,000 objects.
Partial blocks of 100,000 are rounded up. For example, 115,100 excess objects are billed as 200,000 objects (2 blocks).
Billing cycle
Usage is counted monthly. Bills are generated and payments are settled on the 1st of each month. If your usage falls below the minimum charge threshold, no charge applies.
Billing examples
| Scenario | Total objects | Excess objects | Rounded up to | Charge |
|---|---|---|---|---|
| Under free tier | 800,000 | 0 | — | Free |
| Exactly at free tier | 1,000,000 | 0 | — | Free |
| Slightly above free tier | 1,050,000 | 50,000 | 100,000 (1 block) | USD 1 |
| Well above free tier | 1,115,100 | 115,100 | 200,000 (2 blocks) | USD 2 |
Detailed example
On January 1, your data lake contains 100 databases, 10,000 tables, 500,000 partitions, and 5,000 indexes — 515,100 objects in total.
By January 31, you add 600,000 partitions. The total grows to 1,115,100 objects.
On February 1, your bill shows:
-
First 1,000,000 objects: free
-
Excess: 115,100 objects, rounded up to 200,000 (2 blocks) — USD 2
-
On February 1, you receive a bill. The total charge is USD 0 (for the first 1 million free objects) + USD 2 (for the remaining 115,100 objects, which are billed as 200,000 objects) = USD 2.
-
On February 1, USD 2 is deducted from your account for the metadata object storage fees incurred in January.
USD 2 is deducted from your account on February 1 as the January storage fee.
Metadata requests
Metadata requests are create, read, modify, and delete operations on databases, tables, partitions, indexes, and functions in your data lake's metadata.
What counts as a metadata request
Requests are generated when you:
-
Execute DDL statements (such as
CREATE TABLE,ALTER TABLE,DROP PARTITION) in a connected compute engine — E-MapReduce (EMR) Hive, EMR Spark SQL, EMR Presto, or MaxCompute -
Manage metadata objects from the DLF console or via API operations (such as creating, updating, or deleting databases, tables, or partitions)
Pricing
| Usage | Price |
|---|---|
| First 1,000,000 requests/month | Free |
| Above 1,000,000 requests | USD 1 per 1,000,000 requests/month |
-
Billing rule: The first 1 million requests each month are free. For usage that exceeds this free tier, you are charged USD 1 for every 1 million additional requests. Any fraction of 1 million requests is billed as a full 1 million requests.
Partial millions are rounded up to the nearest million. For example, 200,000 excess requests are billed as 1,000,000 requests.
Billing cycle
Usage is counted monthly. Bills are generated and payments are settled on the 1st of each month. If your usage falls below the minimum charge threshold, no charge applies.
Billing examples
| Scenario | Total requests | Excess requests | Rounded up to | Charge |
|---|---|---|---|---|
| Under free tier | 800,000 | 0 | — | Free |
| Exactly at free tier | 1,000,000 | 0 | — | Free |
| Slightly above free tier | 1,200,000 | 200,000 | 1,000,000 | USD 1 |
| Well above free tier | 2,300,000 | 1,300,000 | 2,000,000 | USD 2 |
Detailed example
In January, compute engines including EMR, MaxCompute, and Hologres send 1,200,000 requests to your DLF metadata.
On February 1, your bill shows:
-
First 1,000,000 requests: free
-
Excess: 200,000 requests, rounded up to 1,000,000 — USD 1
-
In January, you made 1.2 million metadata requests from compute engines such as EMR, MaxCompute, and Hologres. The total charge for January is USD 0 (for the first 1 million free requests) + USD 1 (for the remaining 200,000 requests, which are billed as 1 million requests) = USD 1.
-
On February 1, USD 1 is deducted from your account for the metadata request fees incurred in January.
USD 1 is deducted from your account on February 1 as the January request fee.
Data ingestion
Data ingestion extracts data from external sources and loads it into your DLF data lake.
This feature has stopped updating.
Data ingestion tasks themselves are free. However, if you use other services to move data, those services incur their own charges. For example, synchronizing data from ApsaraDB RDS to your data lake requires a Data Transmission Service (DTS) instance, which is billed separately.
Data import jobs are free of charge. However, if a data import job uses other products, additional fees may apply. For example, if you use DTS for an RDS data import job, a DTS instance is created, which incurs DTS fees.
Impacts of overdue payments
When a payment becomes overdue, DLF restricts access in three stages based on how long the payment remains outstanding.
Within 72 hours of an overdue payment
Running tasks and existing data are unaffected, but the following operations are blocked:
-
Create data sources
-
Create data ingestion tasks
-
Create databases and tables
-
Start data ingestion tasks
After 72 hours
DLF further restricts access:
-
Running tasks are stopped
-
Metadata API operations are blocked
After 168 hours
DLF releases your resources:
-
Data sources, data ingestion tasks, and metadata objects (databases, tables, and partitions) in your Alibaba Cloud account are deleted
-
Access to the DLF console is revoked
Deleted data cannot be recovered. Back up important data before a payment becomes overdue.