New Features/Enhancements
- Multiple tables can be inserted concurrently in batch mode by JDBC using parameter binding API
- New component TDinsight is released for system administrator to monitor the status of TDengine cluster through Grafana dashboards
- New component taosAdapter, as a HTTP proxy for TDengine, is released to replace old httpd module embedded in TDengine, the interface is kept as same; more important, taosAdapter can support multiple kinds of data collection agents, like statsd, collectd, tcollector, etc; the data generated or collected by multiple kinds of exportors can be written to TDengine through taosAdapter.
- OpenTSDB writing protocols, including telnet and JSON, are supported by taosAdapter.
- DataX Reader and Writer for TDengine are released, for user to easily migrate from OpenTSDB to TDengine and from TDengine to TDengine.
- Client SDK (taosc) and all connectors in other languages (Java, Go, etc) can connect to multiple TDengine simultaneously.
- Tag of JSON type can be used.
- Regular expression filtering can be used against tag names and values of tags in binary type.
- Multiple scalar functions, ceil, floor, round, etc, are added.
- Configuration can be dynamically adjusted through JDBC and Go connector.
- The interpolation query of the precision timestamp has been enhanced.
Performance Improvements
- Pre-calculation module is redesigned with new storage architecture. The performance improvement is huge in case of querying against massive data, especially in case the data size is tens or hundereds of millions of tuples, the improvement may be 100 times in some cases in which pre-calculation is triggered. However, the improvement is not very obvious if the data size is not big.
- Automatic compression mechanism is introduced for table metadata. The metadata may be restructured according to the data distribution and status of data blocks and the process of restructuring is triggered automatically and it is very fast. Even the system startup speed can be improved by this mechanism.
- Optimization is done for the aggregation and interpolation of grouping query, the performance of some cases can be doubled up.