Docs

Torch Log Service

Product Overview

Product features

This topic introduces the main features of Torch Log Service (TLS).

Data collection

TLS offers various flexible data collection methods including LogCollector, API, SDK, open-source components, and cloud product imports. It supports access to dozens of data sources.

Supports collecting text logs from host, text logs from container, and container standard output logs through LogCollector.

Supports access to Volcengine cloud product logs (TOS, CDN, etc.).

Supports importing data from Kafka, TOS, SLS, Elasticsearch, and other channels.

Supports collecting data from mobile devices.

Data storage

TLS provides efficient and reliable data storage solutions.

Supports infrequent data storage to reduce long-term storage costs.

Supports massive data persistent storage. Provides stable performance, and reading and writing convenience.

Data query and analysis

TLS provides powerful data retrieval and analysis features. This feature facilitates the query and analysis of massive logs.

Supports second-level queries on tens of billions of data entries.

Supports standard SQL92 syntax.

Supports multiple query types, including precise query, fuzzy query, full-text query, field query, and phrase query.

Supports fast analysis, contextual query, log downloads, and index rebuilding.

Visualization

TLS offers rich visual charts and interactive dashboards.

Provides more than 20 types of statistical charts to visualize query and analysis results and support multidimensional analysis needs.

Offers interactive dashboards to build comprehensive data analysis panels.

Supports integration with third-party visualization tools like Grafana.

Data transformation

TLS data transformation feature provides dozens of DSL transformation functions to meet various data processing needs.

Data wrangling: Extracts content and converts formats for fields with disorganized data to obtain structured data.

Data flow: Distributes data to multiple downstream systems for storage or computation so the data can be accessed by different business systems.

Data masking: Masks sensitive information within data.

Data filtering: Filters complex and diverse data and retains only core information that requires focus.

SQL analysis

Scheduled SQL analysis supports regular execution of query and analysis statements and stores the analysis results.

Log scheduled query and analysis: based on the specified time window and schedule, TLS executes query and analysis statements regularly. The analysis results are stored in the designated target log topic.

Log aggregation: Aggregates detailed raw log information into logs with appropriate granularity to reduce storage costs.

Log filtering: Filters raw logs and retains only important fields that need to be concerned about.

Data shipping and consumption

TLS supports real-time shipping and consumption functions.

Provides Kafka protocol interface to connect with downstream ecosystems.

Offers SDK-based log consumption. Supports consumer and consumer group modes.

Supports real-time data dumping and fast streaming to TOS and various heterogeneous storage systems.

Alarm

Offers a comprehensive alarm function, including data monitoring, alarm notifications, and alarm management.

Supports scheduled monitoring of query and analysis results through alarm policy. This enables second-level alarms for abnormal data.

Supports multiple notification channels (DingTalk, Feishu or Lark, WeCom, email, SMS, phone, etc.) for real-time business status notifications.

Allows designing custom notification content by using content syntax, content functions, and content variables.

Trace storage

TLS provides distributed tracing based on the OpenTelemetry protocol and standards.

Supports access, storage, query, and analysis of various trace data.

Supports multiple SDK integration methods for seamless connection with open-source systems.

Last updated: 2025.12.03 15:13:24