This topic describes the release notes of Time Series Database (TSDB). For more information about the version description and the download URL of the TSDB SDK, see Release notes for the SDK.
- Released TSDB for Promethues. The TSDB engine is embedded with the Promethues Remote storage adaptor. It seamlessly integrates with the open source on-premises Promethues engine to quickly monitor Kubernetes clusters. TSDB for Promethues
- Released TSQL, the time series structured query language. TSQL features simpler development, comprehensive coverage of time series computing, support for multi-metric join, and more powerful analysis capabilities for multiple business metrics. For more information, see Quick start.
- Enhanced the protection against dirty data written through the api/truncate operation. For more information, see Clean up data.
- Fixed several stability bugs.
- Fixed the bug that the queried data is incomplete due to index updates.
- Released the MultiValue time series to support downsampling and aggregation of multiple metrics. For more information, see Developer Guide.
- Changed the tag parameter from required to optional. You can write and query data without tags.
- Optimized the efficiency of tag-free queries.
- Optimized timeline queries to improve stability.
- Fixed the bug that the max field does not take effect in the /api/dump_meta operation.
- Supported specification change and resizing of instances from mlarge to 24xlarge.
- Supported auto-renewal management for subscription TSDB instances.
- Added the timestamp parameter to the query/last operation. You can query the business data that is generated before the time specified by the timestamp.
- Supported millisecond-level data reading and writing. You can use the msResolution parameter to control the read precision. If msResolution is set to false (default value), the data is read based on the write precision. If msResolution is set to true, all data is returned based on the millisecond precision.
- Supported MultiValue filtering in the SDK. The MultiValue data written by the SDK can be filtered based on the association of multiple metric values, and all metrics are returned together. For more information, see the SDK documentation.
- Released the time and spatial feature to support writing and filtering of spatial GPS data. You can submit a ticket to test this feature.
- Upgraded the TSDB engine. The 48xlarge and 96xlarge instances support unlimited timeline writing. The instances of other types support a timeline writing capability 10 times the original capability.
- Released the time series insight feature. You can create time series insight events in the TSDB console to visualize queries and analysis of time series data.
- Supported importing demo data for testing with a few clicks. You can use the interactive data query and analysis features without developing any code.
- Supported queries of data by a single timestamp.
- Supported data writing, reading, and downsampling within milliseconds.
- Supported two time units for downsampling: month and year.
- Added two fill policy operators: near and after.
- Fixed the issue that the timeline and disk monitoring data are incorrect in the console.
- Fixed the issue that the timeline meta data can still be queried after timelines are cleaned up.
- Fixed the issue that the type of the read data is inconsistent with the type of the written data.
- Fixed the issue that the writing rate and disk monitoring data are incorrect.
- Released a new edition of the SDK to support writing, query, and analysis of MultiValue data.
- Supported writing of string and Boolean data.
- Provided the last operation to query the latest data point of a time series.
- Fixed the issue that an error occurs when a timeline is deleted in specified conditions.
- Added the monitoring and alerting feature for operations and maintenance.
- Integrated with Alibaba Cloud Realtime Compute.
- Supported integration with the data visualization tool Grafana.
- Supported the change of the billing method of an instance from Pay-As-You-Go to Subscription in the console.
- Supported pagination for data query.
- Added three fill policy operators: linear, previous, and fixed.
- Supported filtering of data points by metric value.
- Supported inverted indexing.
- Supported processing of abnormal query requests (overlimit or overtime requests).
- Optimized integration with the Pipeline aggregate engine.
- Removed restrictions on special characters in tags and metrics.
- Fixed the issue that data cleanup and metric deletion cannot take effect.
- Fixed the issue that the list page cannot be displayed normally in the console.
- Fixed the issue that an error occurs during concurrent queries.
- Optimized the instance stability to ensure service stability for large queries.
- Used the local time zone of the region where the instance is deployed for downsampling by default.
- Fixed the issue that TagValues cannot be displayed completely in query results in the console.
- Provided the getlastdp operation to query the latest time series data.
- Available in the China (Shenzhen) region and provided three zones: A, B, and C.
- Optimized the API operation for timeline deletion to improve the deletion efficiency.
- Supported access from the Internet and access from VPC networks.
- Released the Client for JAVA 1.0.0, which provides the standard API, non-block I/Os, instance protection, automatic batch sending, automatic connection management, and other features.
- Supported deleting millions of timelines in seconds through the API or the console.
- Optimized instance protection by supporting customizing the number of data points per request. This ensures the instance availability.
- Available in the China (Qingdao) and China (Beijing) regions.
- Optimized automated operations and maintenance to support automatic alerting for instance unavailability.
- Fixed the issue of instance unavailability caused by large time ranges in queries.
- Fixed the issue of instance unavailability due to the incompatibility of underlying asynchronous request API operations.
- Integrated with Resource Access Management (RAM) to enable the resource permission management across Alibaba Cloud accounts and RAM users.
- Integrated with CloudMonitor to enable alerting for instance metrics.
- Optimized the instance creation process. The creation progress is displayed in percentage on the Instances page in real time.
- Optimized the instance configuration to support a maximum of 16 tags for each timeline.
- Fixed the issue that the instance name is displayed abnormally if the instance name is not specified during the instance creation.
- Fixed the issue that the curve is displayed abnormally in the instance monitoring chart.
- Fixed the instance creation exceptions caused by placing orders in non-VPC networks.
TSDB V1.0.0 was officially available for public preview. It provides the following features:
Reading and writing of time series data based on OpenTSDB.
Data storage with a high compression ratio. The storage space required by a data point is less than 2 bytes.
Data TTL: You can set a TTL for data. Expired data will be automatically deleted.
Data deletion: You can delete useless or dirty data to release the storage space.
Data query: You can query data on the web client.
Instance monitoring and alerting.
- The TSDB console supports real-time monitoring of three key metrics: writing TPS, total storage timelines, and storage space.
By default, TSDB instances can only be accessed through VPC networks, which enhances instance protection.
You can configure the IP address whitelist to secure your instances and data.