Real-time monitoring and analysis of query execution
Example embodiments relate to providing real-time monitoring and analysis of query execution. In example embodiments, a query plan is obtained for a database query that is scheduled for execution in a database. A query tree is then generated based on the query plan, where the query tree includes operator nodes that are associated corresponding operators in the query plan. At this stage, performance metrics are collected from the database during the execution of the database query. Next, the query tree is displayed with the performance metrics, where a related portion of the performance metrics are displayed in each of the operator nodes.
1. A system for providing real-time monitoring and analysis of query execution, the system comprising of:
a processor to:
obtain a query plan for a database query scheduled for execution in a database;
generate a query tree based on the query plan, wherein the query tree comprises a plurality of operator nodes each of which is associated with a corresponding operator in the query plan;
collect performance metrics from the database, wherein the performance metrics measure performance of the database query during execution of the database query in the database;
display the query tree with the performance metrics, wherein a related portion of the performance metrics is displayed as a graphical representation inside each of the plurality of operator nodes;
as the execution of the database query continues, collect updates to the performance metrics from the database; and
in response to collecting the updates to the performance metrics, update, in real-time, the related portion of the performance metrics displayed as the graphical representation inside each of the plurality of operator nodes based on the updates to the performance metrics.
2. The system of claim 1 , wherein the processor is further to:
display an overview tree of the query plan, wherein the overview tree is used to navigate to a different extent in a display of query tree.
3. The system of claim 1 , wherein the processor is further to:
display a progress indicator to show progress of the execution of the database query, wherein each operator node of the plurality of operator nodes is color-coded based on an execution status of an operator corresponding to the operator node.
4. The system of claim 1 , wherein the database is a parallel database comprising a plurality of database computing devices, and wherein the query plan is executed in parallel on the plurality of database computing devices.
5. The system of claim 1 , wherein the processor is further to:
identify a plurality of critical paths in the query tree based on the performance metrics; and
determine a plurality of tree subsets based on the plurality of critical paths, wherein each of the plurality of tree subsets includes a critical path of the plurality of critical paths and descendent nodes of the critical path from the plurality of operator nodes.
6. The system of claim 5 , wherein the processor is further to:
in response to a user selecting a target subset of the plurality of tree subsets, generate a new subquery for the target subset comprising additional syntax that allows operators in the target subset to be executed in the database; and
display an isolated query tree for the new subquery.
7. The system of claim 1 , wherein the display of a query tree suppresses a low priority node of the plurality of operator nodes based on the related portion of performance metrics.
8. The system of claim 1 , wherein the processor is further to:
identify a plurality of critical paths in the query tree based on resource usage determined from the performance metrics; and
responsive to identifying the plurality of critical paths in the query tree, update a display of operator nodes associated with the plurality of critical paths in the query tree.
9. A method for providing real-time monitoring and analysis of query execution, the method comprising:
obtaining a query plan for a database query scheduled for execution in a database;
generating a query tree based on the query plan, wherein the query tree comprises a plurality of operator nodes each of which is associated with a corresponding operator in the query plan;
collecting performance metrics, in real-time, from the database during the execution of the database query;
displaying the query tree with the performance metrics, in real-time, wherein a related portion of the performance metrics is displayed as a graphical representation inside each of the plurality of operator nodes;
identifying a plurality of critical paths in the query tree based on resource usage determined from the performance metrics; and
responsive to identifying the plurality of critical paths in the query tree, updating a display of operator nodes associated with the plurality of critical paths in the query tree.
10. The method of claim 9 , further comprising:
display a progress indicator to show progress of the execution of the database query, wherein each operator node of the plurality of operator nodes is color-coded based on an execution status of an operator corresponding to the operator node.
11. The method of claim 9 , further comprising:
determining a plurality of tree subsets based on the plurality of critical paths, wherein each of the plurality of tree subsets includes a critical path of the plurality of critical paths and descendent nodes of the critical path from the plurality of operator nodes.
12. The method of claim 11 , further comprising:
in response to a user selecting a target subset of the plurality of tree subsets, generating a new subquery for the target subset comprising additional syntax that allows operators in the target subset to be executed in the database; and
displaying an isolated query tree for the new subquery.
13. The method of claim 9 , wherein a display of query tree suppresses a low priority node of the plurality of operator nodes based on the related portion of performance metrics.
14. The method of claim 9 , further comprising:
as the execution of the database query continues, collecting updates to the performance metrics from the database; and
responsive to receiving the updates to the performance metrics, updating the related portion displayed as the graphical representation inside each of the plurality of operator nodes based on the updates to the performance metrics.
15. A non-transitory machine-readable storage medium encoded with instructions executable by a processor for providing real-time monitoring and analysis of query execution, the machine-readable storage medium comprising instructions to:
obtain a query plan for a database query scheduled for execution in a database, wherein the database is a parallel database comprising a plurality of database computing devices, and wherein the query plan is executed in parallel on the plurality of database computing devices;
generate a query tree based on the query plan, wherein the query tree comprises a plurality of operator nodes each of which is associated with a corresponding operator in the query plan;
collect performance metrics from the plurality of database devices, wherein the performance metrics measure performance of the database query during execution of the database query on the plurality of database devices;
display the query tree with the performance metrics, wherein a related portion of the performance metrics is displayed as a graphical representation inside each of the plurality of operator nodes, and wherein each of the plurality of operator nodes is shown to be associated with at least one of the plurality of database devices;
as the execution of the database query continues, collect updates to the performance metrics from the database; and
in response to collecting the updates to the performance metrics from the database, update, in real-time, the related portion of the performance metrics displayed as the graphical representation inside each of the plurality of operator nodes based on the updates to the performance metrics.
16. The machine-readable storage medium of claim 15 , wherein the machine-readable storage medium further comprises instructions to:
identify a plurality of critical paths in the query tree based on the performance metrics; and
determine a plurality of tree subsets based on the plurality of critical paths, wherein each of the plurality of tree subsets includes a critical path of the plurality of critical paths and descendent nodes of the critical path from the plurality of operator nodes.
17. The machine-readable storage medium of claim 16 , wherein the machine-readable storage medium further comprises instructions to:
in response to a user selecting a target subset of the plurality of tree subsets, generate a new subquery for the target subset comprising additional syntax that allows operators in the target subset to be executed in the database; and
display an isolated query tree for the new subquery.
18. The machine-readable storage medium of claim 15 , wherein the processor is further to:
identify a plurality of critical paths in the query tree based on resource usage determined from the performance metrics; and
responsive to identifying the plurality of critical paths in the query tree, update a display of operator nodes associated with the plurality of critical paths in the query tree.