Version Guide-Time Series Database(TSDB)-阿里云帮助中心

V2.6.13 — April 30, 2021

Fixed issues

IgnoreError mode inconsistency: Fixed behavior inconsistency caused by uncaught exceptions when using the IgnoreError mode.
Empty query result with early startTime: Fixed an issue where no data was returned when startTime was set to an excessively early point in time.

V2.6.12 — April 9, 2021

Design optimization

Write audit logs: Audit logs that record data writes are now available. Because this feature affects write performance, submit a ticket to enable it. Once enabled, you can query audit logs only in the background during O&M.

V2.6.11 — April 2, 2021

New features

roffset and rlimit parameters: Added roffset and rlimit to api/mquery and api/query/mlast. roffset is syntax-compatible with tupleOffset; rlimit is syntax-compatible with tupleLimit.

Fixed issues

Timestamp range validation: Fixed an issue with the valid range of timestamps.
Error on missing metric: Fixed an error that occurred when the queried metric did not exist.

V2.6.10 — February 25, 2021

New features

CloudMonitor metrics: Added storage usage and timeline usage as TSDB instance metrics. Use these metrics to create alert rules in the CloudMonitor console.

Fixed issues

Global limit internal errors: Fixed internal errors that occurred when global limits were applied.
IgnoreError mode not taking effect: Fixed an issue where the IgnoreError mode had no effect.

V2.6.9 — January 27, 2021

Design optimization

Latest value query performance: Improved query performance for fetching the latest values.

Fixed issues

Timeline count after TRUNCATE: Fixed an issue where the timeline count was not reset immediately after a TRUNCATE operation.
Node status after unplanned restart: Fixed an issue where node status was not reset after an unplanned restart during an asynchronous task.

V2.6.7 — January 14, 2021

New features

Field value prefix matching in multi-value queries: The prefix of a field value can now be used as a match condition in multi-value data queries.
Node scale-out for additional editions: Instances in Standard Edition II, Ultimate Edition I, and Ultimate Edition II now support adding nodes during scale-up.
Reduced slow query log blank screen time: The blank screen time of slow query logs has been reduced. For details, see Slow query logs.

Fixed issues

TSQL memory leak: Fixed a memory leak in TSQL.
Error on missing field in multi-value query: Fixed an error that occurred when a queried field did not exist during a multi-value query.

V2.6.6 — January 7, 2021

Download formatted slow query logs: Formatted slow query log content can now be downloaded.

V2.6.5 — December 30, 2020

New features

Global limit on query result size: Applied a global limit to the total number of data points returned per query.

Design optimization

Hint pushdown: Hints can now be pushed down to improve query efficiency.

Fixed issues

Empty query result from null pointer exception: Fixed an issue where a null pointer exception caused query results to be empty.

V2.6.4 — December 3, 2020

New features

Spatial data support (public preview): Spatial data can now be written and queried. This feature is in public preview. Submit a ticket to apply for a free trial.

Fixed issues

Authentication error caused by tsdb user: Fixed an authentication error triggered by the created tsdb user.
Parallel node startup failure: Fixed an issue where nodes could fail to start when multiple nodes started in parallel.

V2.6.3 — October 14, 2020

New features

Query hints: Query hints can now be specified in query criteria. For details, see Query hint description.
Time series forecasting and anomaly detection: Two new analytical features are now supported: time series forecasting and time series forecasting for anomaly detection. For details, see Time series forecasting and Time series forecasting for anomaly detection (abnormaldetect).
Error skipping in batch writes: Certain errors can now be ignored when writing data in batches.

Fixed issues

Empty result with tag key and value filters: Fixed an issue where no data was returned when query conditions included both tag keys and tag values.

V2.6.2 — August 13, 2020

Byte-array data support: Byte-array data can now be written and queried.

V2.6.1 — July 30, 2020

New features

Tuple-based filtering in multi-value queries: Tuples can now be used to filter data in multi-value queries. Specify filter conditions in WHERE clauses.
Reverse queries: Reverse query order is now supported.

Fixed issues

Error on missing tag values in multi-value queries: Fixed an error that occurred when specified tag values did not exist in multi-value data queries.
Out-of-order data not queryable after cache replacement: Fixed an issue where out-of-order data written to existing timelines could not be queried after a cache replacement.
Heap memory issue: Fixed a memory issue on the engine caused by improper heap memory usage.

V2.6.0 — June 29, 2020

New features

INSERT statement support in TSQL: TSQL now supports INSERT statements.

Fixed issues

Connection pool memory issue: Fixed a memory issue caused by an excessive number of connections in a connection pool.
Millisecond data parsing in downsampling expressions: Fixed an issue where millisecond-level data in downsampling expressions could not be parsed.
Error in latest multi-value data queries: Fixed an intermittent error that occurred when querying the latest multi-value data.

V2.5.14 — May 19, 2020

New features

Actual timestamp aggregate functions: Added rmax, rmin, rfirst, and rlast aggregate functions to downsampling expressions. These functions return results with actual timestamp versions.

Fixed issues

Internal cache memory leak: Fixed a memory leak in internal caches.
Unplanned large queries from incorrect expression parsing: Fixed an issue where downsampling expressions were parsed incorrectly, triggering unexpectedly large queries.

V2.5.13 — March 19, 2020

New features

Account management: Account management is now supported. Manage accounts using basic access authentication. For details, see Manage accounts.

Fixed issues

Wildcard tag matching error: Fixed an error that occurred when using wildcards to match tags.

V2.5.12 — January 16, 2020

New features

Prefix and suffix matching filters: Prefix matching and suffix matching can now be used as query filters.

Fixed issues

JSON field misuse in api/query/last: Fixed an issue where the api/query/last operation did not strictly validate JSON fields, causing them to be used incorrectly.

V2.5.11 — November 21, 2019

Fixed issues

Invalid time window with weekly calendar downsampling: Fixed an issue where the time window in a query result was invalid when the downsampling time granularity was set to weeks in calendar form.
Incorrect COUNT result with preDpValue filter: Fixed an issue where the COUNT function returned an invalid result when preDpValue was used to filter values.
delta and rate operator conflict in multi-value model: Fixed an issue where the delta and rate operators were mixed up in the multi-value data model.

V2.5.10 — October 16, 2019

Fixed issues

Invalid TSDB monitoring metric data: Fixed an issue where monitoring metric data in TSDB was invalid.

V2.5.9 — September 3, 2019

New features

delta operator for queries: The delta operator is now supported in queries.

Fixed issues

Tag filter parsing inconsistency: Fixed an issue where a tag filter in the tags field failed to be parsed, while the same filter in the filters field was parsed correctly.

V2.5.8 — August 6, 2019

New features

Median operator for downsampling: The median operator is now supported in downsampling.

Fixed issues

/api/suggest result count not reaching maximum: Fixed an issue where the number of data points returned by /api/suggest could not reach the specified maximum.

V2.5.7 — July 19, 2019

New features

Multi-value data model in TSQL: TSQL now supports the multi-value data model.
New instance specifications: Additional instance specifications are now available.

V2.5.6 — June 27, 2019

New features

Time series data preprocessing: A preprocessing feature for time series data has been added.
Filter-based queries for api/query and api/mquery: Data can now be queried using filter conditions in api/query and api/mquery operations.

V2.5.4 — June 13, 2019

Fixed issues

Content-Type response header backward compatibility: Fixed a backward compatibility issue in the Content-Type response header of api/query.

V2.5.3 — June 10, 2019

API operation changes

Empty result instead of error for missing metric, tag key, or tag value: If the metric, tag key, or tag value specified in a query does not exist, no error is returned. An empty result set is returned instead.

Fixed issues

Prometheus aggregation operator pushdown failure: Fixed query failures that occurred when Prometheus pushed down aggregation operators not supported by TSDB.

V2.5.2 — May 18, 2019

New features

stddev aggregation algorithm: The stddev (standard deviation) aggregation algorithm is now supported.
Paginated results for api/query/mlast in tupleFormat: When the result set of api/query/mlast is in tupleFormat, results can now be paginated.

V2.5.1 — April 26, 2019

Fixed issues

Incorrect data point statistics with details or summary parameter: Fixed an issue where data point statistics were invalid when the details or summary parameter was included in /api/put or /api/mput requests.
JSON parse failure from space in single data point: Fixed an issue where JSON parsing failed when a single data point contained a space.
Null pointer exception in multi-value data model: Fixed a null pointer exception in the multi-value data model.
Timeline count could not be limited: Fixed an issue where the number of timelines in TSDB kept increasing without limit.
Residual schema information after failed timeline deletion: Fixed an issue where some schema information was retained when a timeline failed to be deleted.

V2.5.0 — April 13, 2019

New features

TSQL: TSQL is released. TSQL is a structured query language for TSDB that supports a wide range of computing operations on time series data, including joining multiple metrics and analyzing business metrics. For details, see Quick start for TSQL.
TSDB for Prometheus: Prometheus remote storage support is now built into the TSDB engine, enabling seamless integration with self-managed Prometheus deployments for Kubernetes cluster monitoring. For details, see TSDB for Prometheus.

V2.4.2 — March 20, 2019

Data cleanup with api/truncate: The api/truncate operation can now be used to clean dirty data. For details, see Clean data.
Stability fixes: Several stability issues have been fixed.

V2.4.1 — February 28, 2019

Fixed issues

Data integrity affected by index updates: Fixed an issue where index updates negatively affected the integrity of retrieved data.

V2.4.0 — February 14, 2019

New features

Multi-value data model: The multi-value data model for time series data is now available. Downsampling and aggregation across multiple metrics are supported. For details, see Developer guide.

V2.3.8 — January 24, 2019

New features and optimization

Optional tag parameter for writes: Tags are now optional when writing data. Queries and writes without specified tags are supported.
Tagless query optimization: Queries without specified tags have been optimized.
Timeline query stability: Timeline queries have been optimized for improved stability.

Fixed issues

max field in /api/dump_meta not taking effect: Fixed an issue where the max field in /api/dump_meta requests had no effect.

V2.3.3 — November 27, 2018

Experience optimization

Instance scaling for mid-to-large specs: Instances with specifications from mlarge to 24xlarge now support scale-up, scale-down, scale-out, and scale-in.
Auto-renewal for subscription instances: Auto-renewal is now supported for TSDB instances using the subscription billing method.

V2.3.1 — October 20, 2018

New features

Time-bounded latest value queries: The query/last operation now accepts a time parameter to retrieve the latest data generated before a specified point in time (using data timestamps).
Millisecond read/write granularity: Millisecond-level granularity is now supported for data reads and writes. The msResolution parameter controls read granularity. When set to true, all returned data uses millisecond timestamps. The default value is false, which means the time granularity for data writes is the same as the time granularity for data reads.
Multi-value data point filtering in TSDB SDK: TSDB SDK now supports filtering multi-value data points by combining values across multiple metrics. When filtering criteria are met, all metric values are returned. For details, see the TSDB SDK documentation.
Spatio-temporal support (preview): GPS data can be written, filtered, and analyzed. Submit a ticket to access this feature for testing.

V2.3.0 — October 9, 2018

New features

Increased timeline limits: For 48xlarge and 96xlarge instances, the number of writable timelines is now unlimited. For all other instance specifications, the maximum number of writable timelines has been increased to 10 times the previous limit.

V2.2.8 — August 10, 2018

New features

Single point-in-time data queries: Data at a single point in time can now be queried.
Millisecond-level read, write, and downsampling: Millisecond granularity is now supported for reading, writing, and downsampling data.
Month and year downsampling time units: Month and year have been added as supported time units for downsampling.
near and after fill policies: Two new data filling policies are now available: near and after.

Fixed issues

Invalid time series and disk monitoring data: Fixed an issue where monitoring data for time series and disks was invalid.
Time series metadata persisted after clear: Fixed an issue where time series metadata remained retrievable after the time series were cleared.
Read/write data type inconsistency: Fixed an issue where the type of data read differed from the type written.
Invalid write rate and disk monitoring data: Fixed an issue where monitoring data for write rates and disks was invalid.

V2.2.7 — July 6, 2018

New features

TSDB SDK update: TSDB SDK has been updated to support writing, querying, and analyzing multi-value data points.

V2.2.4 — May 24, 2018

New features

STRING and BOOLEAN data type support: Data of the STRING and BOOLEAN types can now be written.
latest data point query: The last operation is now available to query the latest data point in a time series.

Fixed issues

Timeline deletion not taking effect: Fixed an issue where deleting timelines did not take effect under certain conditions.

V2.2.3 — April 10, 2018

New features

O&M monitoring and alerting: Monitoring and alerting for O&M are now available.
Realtime Compute for Apache Flink integration: TSDB instances can now connect to Realtime Compute for Apache Flink for stream computing.
Grafana integration: TSDB instances can now connect to Grafana for data visualization.
Billing method change in console: The billing method of a TSDB instance can now be changed from pay-as-you-go to subscription in the TSDB console.
Paginated query results: Query results can now be returned page by page.
Data filling policies: Three new filling policies are available: linear (linear interpolation), previous (interpolate using the preceding value), and fixed (interpolate using a fixed value).
Metric value filtering: Data points can now be filtered by metric value.
Inverted index support: Inverted indexes are now supported.
Abnormal query handling: Timed-out requests and requests with invalid parameters are now handled gracefully.

Optimized features

Pipeline aggregation engine access: Access to the pipeline aggregation engine has been optimized.
Special characters in tag and metric names: The restrictions on special characters in tag names and metric names have been removed.

Fixed issues

Data and metric deletion not taking effect: Fixed an issue where operations to delete data or metrics had no effect.
Instances page display error: Fixed an issue where the Instances page in the TSDB console did not display as expected.
Error during concurrent queries: Fixed an error that occurred when running concurrent queries.

V1.2.0 — February 7, 2018

Optimized features

Instance stability under large queries: Improved TSDB instance stability. Fixed an issue where instances became unstable when large queries were run.
Local timezone for downsampled data: By default, downsampled data for an instance is now calculated based on the local time zone of the region where the instance is deployed.

Fixed issues

Incomplete tag values in console: Fixed an issue where tag value queries in the TSDB console returned incomplete results.

V1.1.2 — November 30, 2017

New features

getlastdp operation: The getlastdp operation is now available to query the latest data values of time series.
China (Shenzhen) region: TSDB is now available in the China (Shenzhen) region, with three zones: Zone A, Zone B, and Zone C.

Optimized features

Timeline deletion API: The API operation for deleting timelines has been optimized.

V1.1.1 — November 16, 2017

New features

Internet and VPC access: TSDB instances can now be accessed over the Internet or a virtual private cloud (VPC).

V1.1.0 — November 3, 2017

New features

TSDB Client for Java 1.0.0: Released with standard API operations, non-blocking I/O, instance protection, automated batch delivery, and automated connection management.
Timeline deletion via API or console: Timelines can now be deleted through an API operation or in the TSDB console. Millions of timelines can be deleted within seconds.
Instance protection: The number of data points per request can be specified to protect instance availability.
China (Qingdao) and China (Beijing) regions: TSDB is now available in these regions.

Optimized features

Automated O&M alerts: Alerts are now automatically sent if instances become unavailable.

Fixed issues

Instance unavailability from long query time range: Fixed an issue where a long query time range caused an instance to become unavailable.
Instance unavailability from async API incompatibility: Fixed an issue where incompatible underlying API operations for asynchronous requests caused an instance to become unavailable.

V1.0.2 — September 15, 2017

New features

Resource Access Management (RAM) integration: TSDB can now connect to Resource Access Management (RAM) to manage Alibaba Cloud accounts, RAM users, and instance resource permissions.

V1.0.1 — August 28, 2017

New features

CloudMonitor integration: TSDB can now connect to CloudMonitor. Alert notifications are sent based on monitored instance metrics.

Optimized features

Instance creation progress: The instance creation process has been improved. The Instances page now shows real-time creation progress.
Maximum tags per timeline: The maximum number of tags for a single timeline has been increased to 16.

Fixed issues

Instance name appearing after creation without a name specified: Fixed an issue where an instance name appeared even though none was specified during creation.
Instance monitoring curve not displaying correctly: Fixed an issue where a monitoring curve for an instance did not display as expected.
Instance created without VPC not usable: Fixed an issue where an instance could be created even when no VPCs were available, but could not be used afterward.

V1.0.0 — July 31, 2017

TSDB V1.0.0 is now available in public preview with the following features:

OpenTSDB protocol support: Read and write time series data using the OpenTSDB protocol.
High compression storage: Data is compressed before storage. Each data point requires less than 2 bytes.
Data Management (DMS):
- Data validity: Customize data validity periods. Data is automatically invalidated and deleted after the period ends.
- Data deletion: Automatically delete useless or dirty data to free storage space.
- Data query: Query data on demand through the web UI.
Instance monitoring and alerting: Monitor three core metrics in real time in the TSDB console: write operations per second, total stored timelines, and used storage.
Security:
- TSDB instances connect only over VPCs by default, securing all connections.
- Instance whitelists can be configured to protect your instance and data.