Automated Statistics Management or AutoStats is an autonomous tuning feature that collects statistics on selected tables and columns to improve query optimization and performance. It analyzes logged query workloads and determines the subset of columns that would benefit the most from having statistics collected on them. If the statistics subsequently become stale due to update activity on their underlying data, it automatically refreshes them. As workloads change over time, it removes individual statistics that no longer needed to reduce the overhead of maintaining them.
Users control which databases or tables are managed by AutoStats via new DDL statement syntax (AUTODBA keyword) and can optionally specify a “semi-autonmous” mode (WITH CHECK keyword) that requires approval of tuning recommendations prior to their application. AutoStats tuning is supported on the majority of data storage types offered in VantageCloud Lake including Block Storage, Object File System (OFS), and Native Object Store (NOS). For more information, refer to SQL Data Definition Language and Data Storage.
AutoStats tuning operations run automatically as part of a background service within VantageCloud Lake. This service replaces the Teradata ViewpointViewpoint Stats Manager portlet that controls AutoStats operations within VantageCloud Enterprise. As tables marked AUTODBA are created and loaded with data, AutoStats will begin collecting summary statistics on them and, after observing queries on them, it will recommend additional detailed statistics on columns frequently used in accessing the data. Tuning recommendations made on AUTODBA marked objects will be automatically applied by the system in a fully autonomous fashion. For those objects that include the optional WITH CHECK clause, recommendations are not applied until approved by users. APIs in the form of Stored Procedures allow users to review tuning recommendations which include supporting evidence that describe their anticipated benefit. For more information, see Manual Control and Customization APIs.
Dashboard APIs allow system administrators to monitor AutoStats tuning activity over time. A set of key performance indicators (KPIs) summarize daily tuning activity and its overhead. A separate Audit Trail API reports detailed information about individual tuning actions that occurred during a specified time period and can be used to diagnose issues that may result from AutoStats operations. For more information, see Dashboard Reporting APIs.