Features
Time Series Database (TSDB) is a cloud-native database built for high-throughput time series workloads. It ingests up to 10 million data points per second and returns query results for one million data points in under five seconds.
Data ingestion
Write time series data through HTTP or TSDB Java Client.
Data query
Query data through HTTP, TSDB Java Client, or the TSDB console. The console supports interactive visualization of query results, including data grouping, downsampling, and space aggregation.
Data lifecycle management
TTL
Set a time-to-live (TTL) on your data through the console or API. When data expires, TSDB marks it as invalid and automatically purges it at the scheduled time.
Data cleanup
Delete data by metric through the console, or use the API for more granular control.
Highly compressed storage
TSDB's compression algorithm reduces each data point to 1–2 bytes on average, reducing overall storage usage by 90% and improving ingestion throughput.
Time series analytics
TSDB provides built-in analytics functions for time series data: downsampling, data interpolation, and space aggregation. These functions handle demanding query scenarios.
Monitoring and operations
The built-in operations and maintenance (O&M) system gives you real-time visibility into instance running status, performance metrics, and storage space usage. Configure alerts to detect resource bottlenecks before they affect your workload.
Data and instance security
TSDB provides multiple layers of protection:
VPC-based access: Instances run inside a virtual private cloud (VPC), isolating them from the public internet.
IP address whitelist: Only servers on the allowlist can reach your instance. A server inside a VPC that is not on the whitelist is still blocked.
Triple-copy redundancy: Each data point is stored in three copies to ensure data availability.
Deployment modes
TSDB offers three deployment modes. Choose the one that matches your availability requirements and operational model.
| Mode | Best for | Availability mechanism |
|---|---|---|
| Single-AZ mode | Workloads where cross-zone data pipeline integration is the priority and you need control over the failover policy | TSDB Tunnel service connects all single-AZ clusters; automatic failover strategy is configurable |
| Multi-AZ mode | Mission-critical workloads requiring zero data loss or near-zero downtime | Automatic switch with strong consistent (RPO = 0) or eventual consistent (minimal Recovery Time Objective) options; transaction and synchronous logs replicated across AZs |
| Serverless mode | Variable or unpredictable workloads where agile capacity management and cost efficiency matter | Storage and compute scale automatically with workload; pay-as-you-go billing |
Single-AZ mode
Deploy a cluster in a single availability zone (AZ). The TSDB Tunnel service — a distributed data tunnel — connects all single-AZ clusters and forms a unified data flow. Configure an automatic failover strategy to handle AZ-level outages.
Multi-AZ mode
Deploy a cluster that spans multiple availability zones, connected through one seed. If an outage occurs, TSDB automatically switches without manual intervention. Two recovery options are available:
Strong consistent switch: Recovery Point Objective (RPO) = 0, no data loss
Eventual consistent switch: Minimal Recovery Time Objective (RTO), near-zero downtime
TSDB replicates both transaction logs and synchronous logs across availability zones, enabling automatic data restoration from the distributed log system after an outage.
Serverless mode
Storage and compute capacity scale automatically with your application workload. Pay only for what you use with a pay-as-you-go billing model — no manual capacity planning required.