Table of Contents

Evaluating user experience at peak times with Pega Predictive Diagnostic Cloud

Identify peak usage times and check whether increased load affects your system's performance to immediately detect when the user experience is poor. These findings can help you evaluate whether the system can run your applications.

Use case

uPlusTelco runs a production system with Pega Platform™ applications. Last week, a marketing campaign that focused on the sale of a new mobile device has caused a significant surge in network traffic. The campaign attracted many new customers, who now frequently access uPlusTelco's online services to look for discounted devices. These surges in traffic occur at unpredictable times of day. As an operations manager, you are responsible for ensuring that users have a valuable experience. According to uPlusTelco system administrators, user experience is acceptable if less than 1% of server interactions take longer than one second.

Before you begin

Ensure that you can access Pega Predictive Diagnostic Cloud™ (PDC). For more information, see Getting started with Pega Predictive Diagnostic Cloud.

Evaluating user experience at peak times

Check whether a recent surge in the number of users affected the performance of your system.

  1. Log in to PDC.
    1. Go to Pega Support.
    2. In the Support Quick Links section, click Access Predictive Diagnostic Cloud.
  2. In the header of PDC, in the System list, select the system that hosts the application that operated the sale campaign, for example, upt-prod1.
  3. In the navigation pane, click Usage Viewer.
  4. In the Requestor type list, select Web.
    Web requestors are browser sessions, and their slowness directly affects user experience. A web requestor interaction represents a user that interacts with the application through their web browser, for example, by clicking a button.
  5. In the Application list, select the application that serves the sale campaign, for example, upt-sale2020.
  6. In the Interval list, select Custom time.
  7. In the Date from and Date to fields, select the time period when you expect an increased number of users, and then click Filter.
    For the uPlusTelco campaign, the period to investigate starts at 10:00 on February 8 and ends at 19:00 on February 10.
  8. Analyze the information on the Unique users chart and note the exact time period when the number of users increased the most.
    During the uPlusTelco campaign, the largest spike occurred between 12:00 and 16:00 on the final day of the sale, as in the following example:
    Unique users chart
    "Unique users chart"
    Unique users chart
  9. On the Average response time chart, check whether the response time has increased during the spike of unique users.
    During the uPlusTelco campaign, the average response time increased severely, as in the following example:
    Elevated average response time
    "Elevated average response time"
    Elevated average response time
  10. In the Requestor type list, select Service, and then repeat steps 7 through 9.
    Service requestors are sessions for listeners and for access to Pega Platform from an external client system, such as through a service request. Slow responses to these requestors directly affect the user experience because rendering a screen for a user typically involves multiple server interactions.
    In the uPlusTelco campaign, the number of unique users and the average response time for the Service requestor is very similar to the results for the Web requestor.
  11. In the navigation pane, click System Assessment.
    On the System Assessment landing page, you can analyze response times and interaction volume statistics for your systems.
  12. In the Interval list, select Custom time, and then, in the Date from and Date to fields, select the time period of the spike of unique users that you identified in step 8.
    For the uPlusTelco campaign, select the same period as in step 7.
  13. On the Distribution of healthy and slow interactions chart, compare the percentage of healthy interactions (green) with the percentage of slow interactions (red).
    In the following example, the performance of the upt-prod1 system is not satisfactory because significantly more than 1% of interactions were unhealthy (that is, they took more than one second).
    Distribution of healthy and slow interactions
    "Distribution of healthy and slow interactions"
    Distribution of healthy and slow interactions
  14. On the Average time of healthy and slow interactions chart, for each hour in the set time period, compare the average time that healthy interactions took (green bar) with the average time that slow interactions took (red bar).
    For the uPlusTelco application, the average time that an interaction took was very long for both healthy and slow interactions. This result overlaps with the findings in step 13, and provides another reason why your system's performance might require improvements to process usage spikes generated by future campaigns.

Conclusions

You identified the peak usage times of your system and application during a marketing campaign, and assessed the performance of your system during a spike in usage that correlates with the campaign. You detected unusual behavior that requires further investigation before the system can withstand further such events.

What to do next

Identify the most urgent performance problems in your system. For more information, see Issue identification and research with Pega Predictive Diagnostic Cloud.

Suggest Edit

Have a question? Get answers now.

Visit the Collaboration Center to ask questions, engage in discussions, share ideas, and help others.