In Simple Log Service (SLS), each billable item is billed separately. For example, storing logs incurs storage fees, and collecting logs incurs write traffic fees. The following sections describe the billable items and payment methods of the pay-by-feature billing model.
Notes
-
The pay-by-feature model of Log Service supports collecting logs and time series data. Separate pricing applies to storage space and index building for time series data. However, the data processing fee, data shipping fee, read/write traffic fee, and request count fee are the same as for logs.
-
You can view metrics such as write traffic, read traffic, read/write count, data processing traffic, data shipping traffic, and storage usage for the previous day in the Log Service console.
These metrics are updated daily and are not available in real time.
-
During log collection, Log Service automatically compresses log data to 10% to 20% of its original size.
What is an OCU?
Observability Capacity Unit (OCU) is the billing unit for Alibaba Cloud-native Observability. OCU usage is calculated automatically based on hourly resource consumption. On the Alibaba Cloud China Website, a public cloud OCU is priced at CNY 0.15 per unit.
Billing for the compute-optimized features of SLS is transitioning to an OCU-based model that charges based on actual compute resource consumption. In CPU scenarios, one OCU is equivalent to approximately 0.5 CPU cores, 2 GB of memory, and 3,000 IOPS. SLS calculates separate OCU values based on the consumption of CPU cores, memory, and IOPS. The highest of these three values determines the final OCU count for billing.
In GPU scenarios, one OCU is equivalent to approximately 1/60 of the computing power of an A10 card. The final OCU value is based on the consumed GPU computing power.
For example, a computation job that consumes 1 CPU core, 2 GB of memory, and 3,000 IOPS uses 2 OCUs. In a one-hour metering period, processing 1 GB of data consumes approximately 1/3 of an OCU for an ingest processor, 1/3 of an OCU for data transformation, and 0.3 OCUs for rule-based consumption.
For example, a vectorization job that uses the SLS vector index feature and consumes 1/60 of an A10 card's computing power over a one-hour metering period uses 1 OCU.
Billable items for pay-by-feature
The following table describes the billable items of the pay-by-feature model for Log Service. For detailed pricing information, see Pricing.
|
Billable item |
Description |
Billing method |
Free quota |
|
storage space - hot storage for logs |
This includes the storage size of compressed logs and the indexes created from uncompressed logs. The indexes can be a full-text index, field index, or scan index. For example, if your raw logs are 1 GB in size, after you upload the logs to Log Service and configure indexes for two target fields, the data is compressed to 20% of its original size and the index data for the two fields is 0.5 GB. In this case, the total storage space is 0.2 GB + 0.5 GB = 0.7 GB. |
|
500 MB/month Important
The free quota is available only if you do not use a resource plan to offset fees. |
|
storage space - IA storage for logs |
After you enable intelligent tiered storage, logs are converted to Infrequent Access (IA) storage (formerly cold storage) when their storage duration exceeds the configured Hot Data Retention Period. You are then charged for the IA storage space. This storage space includes the volume of compressed logs and the indexes created from uncompressed logs. For example, if your raw logs are 1 GB in size, after you upload the logs to Log Service and configure indexes for two target fields, the data is compressed to 20% of its original size and the index data for the two fields is 0.5 GB. In this case, the IA storage space is 0.2 GB + 0.5 GB = 0.7 GB. |
|
None |
|
storage space - archive storage for logs |
After you enable intelligent tiered storage, logs are moved to archive storage when their configured Hot Data Retention Period or IA-tier data retention period ends. You are then charged for the archive storage space used. This space includes the volume of compressed logs and the index data generated from the uncompressed logs. For example, if your raw logs are 1 GB in size, after you upload the logs to Log Service and configure indexes for two target fields, the data is compressed to 20% of its original size and the index data for the two fields is 0.5 GB. In this case, the archive storage space is 0.2 GB + 0.5 GB = 0.7 GB. |
|
None |
|
storage space - time series storage |
This includes the storage size of compressed time series data and the indexes created from uncompressed time series data. For example, if your raw time series data is 1 GB in size, after you upload the data, Log Service automatically builds an index of 1 GB. If the data is compressed to 20% of its original size, the total storage space consumed is 0.2 GB + 1 GB = 1.2 GB. |
|
None |
|
read and write traffic |
Read and write traffic consists of two components:
Note
The consumption preview feature in the Log Service console generates a small amount of read traffic. |
|
500 MB/month Important
The free quota is available only if you do not use a resource plan to offset fees. |
|
index traffic - log indexing |
Index traffic is calculated based on the volume of index data generated when uncompressed logs are indexed or reindexed. Indexes are created for specific fields. The amount of index traffic depends on the indexed fields and the length of their values.
Important
|
|
500 MB/month Important
The free quota is available only if you do not use a resource plan to offset fees. |
|
index traffic - log indexing - query type |
Index traffic is calculated based on the volume of index data generated when uncompressed logs are indexed or reindexed. Indexes are created for specific fields. The amount of index traffic depends on the indexed fields and the length of their values.
Important
|
|
None |
|
index traffic - scan index |
Index traffic is calculated based on the volume of index data generated when a scan index is built or reindexed for uncompressed logs. The amount of index traffic depends on the log fields and the length of their values. If a field has both a full-text or field index and a scan index, only the traffic for the full-text or field index is billed. Creating a scan index generates index storage data equal to 20% of the raw log volume.
|
|
None |
|
index traffic - time series index |
Index traffic is calculated based on the volume of index data generated when an index is built for uncompressed time series data. Indexes are created for specific fields. The amount of index traffic depends on the indexed fields and the length of their values. When you upload time series data, Log Service automatically builds an index for the data. For example, if you write 1 GB of raw time series data to Log Service, 1 GB of index traffic is generated. |
|
None |
|
public network read traffic |
Public network read traffic is generated when you pull data from a Log Service public endpoint. Traffic is calculated based on the compressed data volume. |
|
None |
|
scan traffic |
Scan traffic is calculated based on the volume of uncompressed data that is scanned. |
|
None |
|
data transformation |
Fees are calculated based on the volume of uncompressed data transformed. If you transform data across regions, public network read traffic is generated. This traffic is calculated based on the compressed data volume. |
|
None |
|
data shipping |
Data shipping fees are calculated based on the volume of uncompressed data that is shipped to destinations such as OSS, MaxCompute, and AnalyticDB for MySQL. Note
Shipping data to AnalyticDB for MySQL also incurs fees for read and write requests and read traffic. |
|
None |
|
read and write requests |
Note
Both successful and failed read and write requests are counted. |
|
1 million requests/month Important
The free quota is available only if you do not use a resource plan to offset fees. |
|
voice call |
You are charged per voice call for alert notifications. Note
|
|
None |
|
text message notification |
Alert notifications sent by text message are charged per notification. Note
Some carriers may split long text messages (for example, over 70 characters) into two messages. You may receive two messages, but Log Service charges for only one notification. |
|
None |
|
active shard lease |
Fees are calculated based on the number of shards in the read/write state. Charges for shards stop the day after they are merged or split. For example, if you merge three active read/write shards into one, you are still charged for three shards on that day. On the next day, you are charged for one shard. Note
|
|
31 shard-days/month Important
The free quota is available only if you do not use a resource plan to offset fees. |
|
dedicated SQL |
Fees are calculated based on the actual CPU time used for SQL analysis. The unit is the core-hour, which represents the work done by one CPU core running for one hour. For more information, see Dedicated SQL billing examples. |
|
None |
|
Rule-based consumption |
You are billed based on the resources consumed to process data by using rule-based consumption. The billable unit is OCU. In a one-hour metering period, rule-based consumption consumes approximately 0.3 OCUs to process 1 GB of data. |
|
None |
|
Data transformation (new version) |
You are billed based on the resources consumed to process data by using the new version of data transformation. The billable unit is OCU. In a one-hour metering period, the new version of data transformation consumes approximately one-third of an OCU to process 1 GB of data. |
|
None |
|
transfer acceleration |
You are billed based on the inbound and outbound traffic generated through a transfer acceleration endpoint. Transfer acceleration is billed based on the actual amount of data transferred. For example, if data is compressed during upload, traffic is calculated based on the compressed size. For more information, see Manage Projects. |
|
None |
|
ingest processor |
An ingest processor processes log data before it is written to a Logstore. For example, you can filter data, extract fields, extend fields, and mask data. Fees are based on the resources consumed for data processing. The billing unit is OCU. In a one-hour metering period, an ingest processor consumes approximately one-third of an OCU to process 1 GB of data. |
|
None |
|
AI computing |
You are billed based on the resources consumed when you use vector indexes to call an embedding model to vectorize data or call a semantic enrichment LLM function to enrich data semantics. The billable unit is OCU. In a one-hour metering period: Calling the Calling the |
|
None |