View performance insights

After connecting your Data Center instance to Portfolio insights and completing the prerequisites, we’ll start to generate performance insights. Once the connection is established, it will take at least one hour before initial performance insights can be generated

How it works

Performance insights uses the Apdex (Application Performance Index) standard to measure how satisfied users are with the response times of key experiences in your Data Center instance. Apdex converts raw response time data into a single, easy-to-understand score that reflects real user satisfaction.

Apdex scoring

Every time a user loads a page or performs an action, the response time is placed into one of three buckets:

The Apdex score is then calculated as:

Apdex = (Satisfied count + 0.5 × Tolerating count) ÷ Total count

The result is scaled to a value between 0 and 100, where 100 means all users were satisfied and 0 means all users were frustrated.

For more details, see Monitoring Data Center performance with Apdex.

Data collection and aggregation

Performance data is collected on your Data Center instance every 2 minutes using browser metrics. The aggregated results are then transferred to Portfolio insights in the cloud and displayed on the dashboard every hour, so results always appear with a slight delay. The Apdex score is calculated per experience (for example, issue view or board view), giving you granular visibility into which parts of your instance are performing well and which need attention.

How recommendations work

While Apdex provides the headline performance score, recommendations are driven by a deeper analysis using Ready for User (RFU) metrics. We look for a combination of signals, such as degraded page load times, elevated database query times, or failing health checks that need to occur together before a recommendation is shown. This increases our confidence that a recommendation is relevant and actionable for your specific instance.

Before you begin

• Make sure you’ve completed the required prerequisites to establish the connection

View performance insights

To view performance insights:

1. Under Portfolio insights, select the Instance health page.

2. Select any Jira or Confluence Data Center instance.

3. In the right-hand panel, select View insights. You’ll be moved to the instance view where you can switch between different tabs, such as Optimization or Performance.

4. Select the Performance tab. You’ll see the performance dashboard with key experience response times and response time trends.

Experience health

Health score for key experiences includes data from the last recorded hour. The graph below the score includes data from the last 24 hours.

Preview
Timeframe	The results show the last recorded hour, which is also included below the specific time. The graphs show aggregated results for the last 24 hours.
Data aggregation	The data is collected on your instance every 2 minutes, but the aggregated results are transferred to cloud and displayed on the dashboard every hour, that’s why you’ll always see them with a delay.
Data mutability	The data shown is confirmed and won’t change.

Jira experiences

Details of what Jira experiences are shown.

Confluence experiences

Details of what Confluence experiences are shown.

Health trend

Health trends allow you to view historical scores in either 24-hour or 7-day view.

24-hour trend

The 24-hour trend allows you to distinguish spikes from trends, shows impact of peak usage times, and helps spot daily patterns. It’s the most common view for regular monitoring.

Preview
Timeframe	Last 24 hours.
Data aggregation	The data is aggregated every 10 minutes, so the complete 24-hour view will include 144 data points.
Data mutability	The view shows historical and confirmed data that will not change.

7-day trend

The 7-day trend allows you to see weekly patterns and clear differences between weekday and weekend performance, and helps validate any performance changes.

Preview
Timeframe	Last 7 days.
Data aggregation	The data is aggregated every 1 hour, so the complete 7-day view will include 168 data points.
Data mutability	The view shows historical and confirmed data that will not change like in the case of an hourly view.

Recommendations to improve your performance (Jira only)

We’ll keep analyzing your performance data and symptoms specific to your instance and show recommendations on how you can improve performance if it’s currently degraded.

How it works

Recommendations are always related to some issues detected in your instance. We look for a set of signals that need to occur together, which increases our confidence that a recommendation is right for your instance and actually solves a performance problem you’re experiencing.

To give you an example, here’s a set of signals from the Review Marketplace apps with high database usage recommendation:

• At least one of the key experiences, such as issue view, shows degraded page load times (RFU)

• The db.core.executionTime metric for a Marketplace app was degraded over the same time window when RFU degraded, compared to previous weeks

• The db.core.executionTime metric is degraded for fewer than 5 apps. This number makes it more likely that the issue is actually related to these apps and isn’t a broader system problem.

When these signals occur together, we have a high confidence that specific Marketplace apps and their database usage contribute to current performance issues. We’ll then show a recommendation, suggesting to review the identified apps and follow actions to resolve the issue.

Initial data analysis

To show recommendations, we first need to establish your performance benchmarks, which might take up to 4 weeks. Most of the recommendations will only appear after this time.

Sample recommendation with details

Here’s a sample Review Marketplace apps with high database usage recommendation, with general details explained.

Description

Every recommendation starts with a description of what we have detected and how it impacts your performance. You’ll also see link for more details about the logic we used to display it.

Signals

In the signals section, you’ll find the most important symptoms we used to detect and issue and show this recommendation. If it was shown multiple times, you’ll be able to switch between the latest, last 24 hours, and last 7 days filters to see historical data.

Degraded experiences

This sections shows which key experiences were degraded during the same time window a recommendation was shown. This is one of the key signals that we use to show recommendations only when your performance is actually degraded.

Identified apps

This section typically shows items related to the detected issue, in this case specific Marketplace apps that you should review. For other recommendations, you’ll see different data here, such as other identified entities or configuration details. If there’s a percentage value available, it means how often a specific entity was involved.

Frequency

The frequency chart is availably only in the 24 hour and 7 day filters. It shows how often the recommendation was triggered over time. It should help you understand when the related issue occurs most often, for example Marketplace apps running automations during a specific day of the week so you can troubleshoot the problem more easily.

Recommended actions

Every recommendation ends with an action on how to resolve the related issue, for example archive issues to reduce the index size.

All available recommendations

There are differences in what we show in different recommendations, but they follow a similar pattern. If you’re interested in what issues we currently detect, you can view all available recommendations below:

Jira performance recommendations