--- title: Compare datasets in Instance Observer description: Instance Observer help you understand the impact of changes, troubleshoot issues, and understand performance patterns over time. locale: en-US canonical_url: https://www.servicenow.com/docs/r/australia/impact/io-compare-datasets-vid-tut.html release: australia topic_type: task last_updated: "2026-06-24" reading_time_minutes: 2 keywords: [Instance Observer, compare datasets, analytics, performance metrics, observability, server response time] breadcrumb: [IO analytics, Monitor instance performance, Platform Health, Using Impact, Impact] --- # Compare datasets in Instance Observer Instance Observer help you understand the impact of changes, troubleshoot issues, and understand performance patterns over time. ## Before you begin Role required: admin ## About this task Use the Compare datasets tool to compare high-level metrics across time. You can select from different dimensions for each dataset. For example, in the Average server response time, you can pick two days for about a three-hour period. Then compare the overlays and analyze differences in performance metrics between the two time periods. This is useful to understand the impact of deployments, configuration changes, or identify performance degradation over time. ## Procedure 1. Navigate to **All** > **Impact** > **Platform Health** > **Monitor**. 2. Log in to **Instance Observer** and navigate to **Analytics** > **Compare Datasets**. 3. In the **Instance** field, select the instance you want to analyze. For example, select `observability` to analyze the observability instance. 4. In the **between** field, select the time range type for your comparison. The default option is Date/Time, which allows you to specify exact date and time ranges for both samples. 5. In the **Compare** field, select the metric you want to compare. For example, select Average Server Response Time to compare server performance between two time periods. 6. In the **Sample 1** field, specify the date and time range for the first dataset. For example, select `May 10, 2026 05:58 AM - May 10, 2026 08:58 PM` to analyze data from May 10. \[Omitted image "io-compare-dataset-server.png"\] Alt text: Compare datasets for average server response time. 7. In the **Sample 2** field, specify the date and time range for the second dataset. For example, select `May 12, 2026 05:58 AM - May 12, 2026 08:58 PM` to compare against data from May 12. 8. Select **Get Snapshot** to generate the comparison. The system displays a comparison of the selected metric between the two time periods, including: - A summary table showing Mean and Median values for both samples - A line chart visualizing the metric trends over time for both samples - Moving average lines to help identify patterns and trends You can use the **Raw** and **Avg** toggle buttons to switch between viewing raw data and averaged data in the chart. 9. Select the **Compare** field and select your preferred metric to compare the datasets. - **Average session summary enduser**: The average time span of a user when they log in until their session ends. - **Average session summary loggedIn**: Total session counts for non-active users. - **SQL Response Time**: The smallest unit of measurement to track SQL per SQL statement. Tracks the average time \(in ms\) for SQL queries to execute against the ServiceNow database If it deviates beyond 2.5 milliseconds, select Query Patterns. Then zoom into 15 minutes on each side of the spike to analyze the graph and the dataset to understand the cause of the issue. 10. Select **Releases** option in the **between** list to compare the date ranges between two upgrades. 1. Select the releases from the two lists of the **Select releases to generate dates for comparison**. 2. Select **Get Snapshot**. For one period, it picks the dates for you between the two releases that you're comparing in the **Using:** section. When you select the \[Omitted image "inline-data-vis-calendar-days.png"\] Alt text: calendar icon., the dates are selected for you to compare the datasets. **Parent Topic:**[IO analytics](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/impact/io-analytics.md)