Trino Queries Details
The Trino Query Details page provides a comprehensive view of each executed Trino query, including the SQL statement, execution metrics, resource usage, and the query execution plan. This helps you analyze query behavior, performance bottlenecks, and system utilization.
Query Overview
In the Summary Panel, you can view high-level details about the query.
Metric | Description |
User | The user who submitted the query (e.g., root). |
State | The final query state (e.g., FINISHED, FAILED, CANCELED). |
Transaction ID | Unique transaction identifier associated with the query. |
of Stages | Number of execution stages in the query plan. |
Duration | Total time taken to complete the query. |
Created Time | Timestamp when the query was created. |
Start Time | Timestamp when the query started execution. |
End Time | Timestamp when the query finished execution. |
Query Text and Catalog Information
This section shows the actual SQL query and the metadata of the involved tables:
- Query: The submitted SQL text.
- Catalog, Schema, Table: Location of referenced tables.
- Authorization: User role used for access.
- Columns: Lists the column(s) involved in the query.
Stage Execution Metrics
Each query is broken down into stages, which represent different parts of the query execution. Each stage includes the following details.
Field | Description |
Stage ID | Unique identifier of the stage. |
Parent ID | Indicates the parent stage in the DAG (if any). |
Total Tasks | Number of parallel tasks run in this stage. |
Failed Tasks | Number of tasks that failed during execution. |
Completed Tasks | Number of tasks that finished successfully. |
CPU Time | Total CPU time consumed by the tasks. |
Completed Drivers | Number of drivers (executors) that completed. |
Peak User Memory | Maximum memory used by a driver for that stage. |
Total Scheduled Time | Total time scheduled across all drivers. |
Blocked Time | Time spent waiting (e.g., for I/O, resources). |
Output Data Size | The total amount of data output by the stage. |
User/Memory Reservation | Reserved memory amount per stage/driver. |
You can expand a stage for detailed tasks and operator-level metrics.
Task Stats Panel (Per Stage)
This displays per-task execution details within a selected stage.
Metric | Description |
Task ID | Unique identifier for the task. |
State | Status of the task (e.g., FINISHED). |
Output Data Size | The amount of data produced. |
Node ID | The specific worker node executing the task. |
Peak User Memory Reservation | |
Garbage Collection (GC) Time | Time spent on Java GC. |
Elapsed Time | Time taken by the task. |
CPU Time | The time CPU consumed for processing. |
Start, Create, End Time | Execution timestamps. |
Operator Stats (Per Task)
This shows the breakdown of operators used within a task.
Metric | Description |
Operator ID and Type | Identifies execution units like ScanFilter, PartitionedOutput. |
Output Data Size | Data emitted by the operator. |
Memory Reservation | Peak Total and User Memory used. |
Plan Node ID | Maps the operator back to the DAG node. |
Total Drivers | Number of drivers executing that operator. |
Query Plan and DAG
The Query Plan and DAG section visualizes how Trino executes the query across its distributed architecture.
DAG (Directed Acyclic Graph) Tab
The DAG tab shows how the query is broken into logical plan nodes, including how data flows between stages and operations. It’s useful for identifying the execution strategy and pinpointing any inefficiencies.
Plan Tab
The Plan tab provides low-level, structured details about the query execution plan as interpreted by Trino. Unlike the DAG view (which is visual), this tab reveals the actual JSON metadata behind each node in the plan.