When queries run slowly or fail, diagnosing the root cause requires access to historical execution data. The Historical Slow Query feature in HoloWeb lets you filter, visualize, and drill into slow and failed query logs so you can identify and resolve performance bottlenecks. With it, you can:
-
Filter historical slow query logs by database, table, user, query type, and time range
-
Spot query volume spikes and failure patterns on a visual trend chart
-
Drill down into individual queries to view execution plans and timing details
-
Analyze concurrent query execution with a Gantt chart to identify performance bottlenecks
Limits
-
Hologres V0.10 or later is required. To check your instance version, go to the instance details page in the Hologres management console. To upgrade, see Common upgrade preparation errors or join the Hologres DingTalk group for support. For more information, see How do I get more online support?.
-
Each search returns a maximum of 2,000 slow query logs.
-
HoloWeb shows historical slow query logs from the last seven days only.
View slow query logs
Prerequisites
Before you begin, make sure that:
-
Your Hologres instance is V0.10 or later
-
You have access to the HoloWeb console
Steps
-
Log on to the HoloWeb console. In the top navigation bar, click Diagnostics and Optimization.
-
In the left navigation pane, click Historical Slow Query.
-
On the Historical Slow Query page, configure the filter parameters.
Parameter Required Description Instance Name Yes The instance to query. The current instance is selected by default. Database No The database to query. If you do not have the required permissions, only logs related to your account are visible. Table Name No Filter logs to a specific table. If you do not have the required permissions, only logs related to your account are visible. User No Filter by Alibaba Cloud account to narrow results to a specific user's queries. Number of rows to query Yes The number of logs to return. The maximum is 2,000. Running time No The minimum query duration to include. By default, queries running longer than 1 second are collected. Increase this threshold to focus on the slowest queries. Graph dimension No Select Slow Query, Failed Query, or both to control what the trend chart displays. Use this to distinguish between queries that ran slowly and queries that failed entirely. Query No Search for a specific SQL statement. Fuzzy match using the percent sign (%) for table names is supported. Type No Filter by operation type, such as DDL or DML. Use this to isolate write operations from read operations when debugging. Query ID No Look up a specific query by its ID. Use this when a user reports a specific failed query and provides the ID. Compute group Yes The compute resources to query. Time range Yes The time range for the query results. Defaults to the last 10 minutes. You can select up to the last seven days. -
Click Search.
View query trend analysis and query list
Query trend analysis
The trend chart gives you two ways to inspect query activity:
-
View queries around a specific point in time — click a point on the chart to display the trend for the 10 seconds before and after the selected point.
-
View queries within a time range — drag to select a range on the chart to see the trend analysis for that period.
Query list
The query list shows the following columns by default.
| Column | Description |
|---|---|
| Query ID | The unique identifier for the query. |
| Database | The database the query ran against. |
| User Name | The user account that executed the SQL statement. |
| Type | The operation type of the SQL statement. |
| Duration | The total query run time, including optimizer execution time, query start time, and the time to return results. |
| Start Time | The time the query started. |
| Status | The current status of the query. |
| Query | Click to view the query details. |
| Engine Type | The execution engine used for the query. |
| Query Plan | The execution plan for the query. |
| Operation | Three actions are available: Details displays detailed information about the query to help with optimization; the SQL editor icon opens the query in a new editor window; the Query Insight Analysis icon navigates to Query Insight Analysis. |
Add more columns — if the default columns are not enough, click Customize Columns in the upper-right corner of the query list and select the additional fields to display.
View the Gantt chart — to analyze concurrent query execution, click View Gantt Chart in the upper-right corner. The Gantt chart displays overlapping query timelines, making it easier to pinpoint performance bottlenecks in high-concurrency scenarios.
