This topic introduces the billing examples for pay-by-feature billing mode of Torch Log Service (TLS).
TLS pay-by-feature billing mode provides flexible billing by measuring and charging separately for features such as data ingestion, indexing, storage, query, data transformation, and log shipping. Its core value is to help you accurately calculate the costs of each feature module for refined cost management. This mode is suitable for scenarios with complex log processing pipelines and diverse feature usage.
Notice
- For illustrative purpose, all data is calculated for a complete billing cycle. Billing case results are for reference only. For details, see the actual bill.
- The data volume conversion rule is 1 TiB = 1024 GiB.
In the China 2 (Beijing) region, a pay-by-feature log topic is created to collect logs from Website A. Five TiB (5120 GiB) of uncompressed raw logs is written into this log daily. Details are as follows:
- Shards: A total of 102 active read/write shards (including 2 free daily read/write shards, i.e., the billing shards count is 100).
- Data writing: Assuming a raw log compression rate of 20% (write traffic/storage volume calculated based on compressed data volume, i.e., 1024 GiB generated per day).
- Index: Index enabled for all fields (index traffic and storage calculated based on uncompressed data volume, generating 5120 GiB daily).
- API calls: 8.5 billion (8500 million) chargeable API requests daily via LogCollector, Open API, or SDK.
For demonstration only. Actual prices are subject to official website pricing, orders, and bills.
The following cases display the daily cost of one log topic at different stages (not the sum of five days). Each use case adds a new feature based on the previous one.
Website A writes 5 TiB of raw logs in one day, builds indexes for all fields, initiates 8.5 billion billing requests, and configures 102 read/write shards.
Website A continuously writes 5 TiB of raw logs daily. The log topic is retained for 75 days, with the most recent 10 days in standard storage, days 11–35 in infrequent storage, and days 36–75 in archive storage.
Once reaching a steady state, such as on day 76, the daily costs are as follows (each day, 1024 GiB of logs enter storage—after compression, index size is 5120 GiB—uncompressed):
Notice
- If all 75 days of data are stored in standard storage, the storage cost will be CNY 5299.20 per day. If you use the lifecycle above, the cost will be CNY 2089.96 per day.
- This comparison only shows the difference in unit price between storage tiers and does not include potential limitations and recovery requirements when using infrequent access or archived data.
Based on Case 2, 6144 GiB (6 TiB) of data is consumed over the internal network, and 1300 GiB (about 1.27 TiB) is consumed over the public network on that day.
Notice
- The data in this case represents down‑link metered values actually collected by TLS. Traffic is counted repeatedly when the same set of data is consumed multiple times.
- For example, when consumption requests are already included in the fixed daily 8.5 billion billing requests, no additional API fee is charged.
- Internal and external network read traffic is determined by the access domain and network path, and the same outbound traffic will not be billed simultaneously.
- Queries via the console or query-and-analysis API queries do not generate read traffic. Corresponding read traffic is generated only when you download logs over internal/external networks or consume logs via programs.
Case 4 Data transformation logs
Building on Case 3, for the log topic located in China 2 (Beijing), 7168 GiB (7 TiB) of uncompressed logs were processed that day, and the result was written into another pay-by-feature log topic in China 2 (Shanghai). TLS reported a compressed cross-region transfer volume of 1300 GiB for the source log topic.
Notice
- In pay-by-feature mode, data transformation traffic is calculated based on the actual volume processed and does not qualify for the 100 TiB monthly quota.
- This case does not include the costs incurred by the target log topic, such as write traffic, indexing traffic, log and index storage costs.
Case 5 Data shipping logs
Building on Case 4, Website A shipped 20480 GiB (20 TiB) of data in JSON format or to Kafka via the basic edition, and 5120 GiB (5 TiB) of data in Parquet format via the advanced edition on the same day. Neither of the shipping tasks enabled file compression.
Notice
- For pay-by-feature mode, basic log shipping does not receive the 100 TiB monthly quota, and advanced log shipping does not receive the 10 TiB monthly quota. Both types are billed based on actual usage.
- This case does not include the cost of target products, such as storage, request, or traffic fees after shipping to targets like TOS or Kafka.