Time Series Database (TSDB) supports efficient reading and writing of time series data. The response time for reading one million data points is less than five seconds. A maximum of 10 million data points can be written per second.
TSDB allows you to write data through HTTP or TSDB Java Client.
TSDB allows you to query data through HTTP, TSDB Java Client, or the TSDB console. The console-based data query feature supports visualization of data grouping, downsampling, and space aggregation.
Data TTL setting
You can set the TTL for data through the console or the API. After the data TTL is set, the system marks the data as invalid when it expires, and automatically cleans it up at the specified time.
You can clean up data based on metrics in the console or use the API to clean up data more flexibly.
Highly compressed storage
TSDB uses the efficient data compression technology, which reduces the average storage space required for a data point to one to two bytes. This not only reduces overall storage usage by 90%, but also increases the data writing speed.
Computing capability of time series data
TSDB provides professional and comprehensive computing functions for time series data, and supports downsampling, data interpolation, and space aggregation. This can meet the requirements of various complicated data query scenarios.
Monitoring and O&M
TSDB provides an instance O&M system. This system allows you to monitor the instance running status, performance metrics, and storage space usage in real time. You can also configure the alerting mechanism to detect resource bottlenecks at the earliest opportunity.
Data and instance security
TSDB provides the following solutions to protect your data and instances:
Provides the VPC-based instance access mode to fully secure the access to your instance.
Provides the IP address whitelist. Only the servers on the whitelist are allowed to access your instance, which further secures the access to your instance and data. If a server is deployed in a virtual private cloud (VPC) but not included in the whitelist, the server cannot access your instance.
In TSDB, data is stored in three copies to fully ensure data availability.
Three deployed patterns
TSDB is a cloud-native database, which is deployed on availabilities zones across the world. It supports three kinds of deployed modes: Single AZ mode, Multi-AZ mode, and Serverless mode.
Single AZ mode: User can select a specific AZ within one region to build a cluster, and the TSDB Tunnel service, a distributed data tunnel, is provided to connect all single-zone clusters and form data flow. An automatic failover strategy supports to be customized to handle the AZ-level outage.
Multi-AZ mode: User can select more than one AZ to build a cluster, and connect the multi-AZ cluster through one seed. Once the outage happens, both strong consistent switch (0 RPO) and eventual consistent switch (Extremely small RTO) are available, without any of the user intervention. In the multi-AZ mode, TSDB supports storing TSDB tables’ transaction logs and synchronous logs between multiple AZs. Thus, TSDB can restore data from the distributed log system and recover it automatically once the outage occurs.
Serverless mode: Serverless mode can provision storage and compute capabilities according to application workload, and it enables cloud tenant to buy database service by pay-as-you-go mode and provides a much agile capacity planning method.