The eGenie - AI Assistant Console

Clicking the eG AI Assistant button will open Figure 1 where you will be offered in-depth insights into the measure for which alerts were raised.

Figure 1 : The eGenie - AI Assistant console

The eGenie - AI Assistant console consists of the following sections:

  1. Explanation: This section, as shown in Figure 1, provides a detailed guide explaining what the alert is, the factors contributing to it, its impact, and ways to address the issue. Administrators can use these insights into identify the pain points and suppress the alert within a short span of time.

  2. Analysis: This section (see Figure 2) outlines the potential consequences that may arise due to the alert raised for the measure. This section helps administrators thoroughly investigate the issue and arrive at a viable solution.

    Figure 2 : The Analysis section

  3. Anomaly Detection: If the Would like to enable detailed insights? flag in the AI ASSISTANT SETTINGS page is set to Yes, then, eG Enterprise automatically uploads the Measure Graph of the problematic measure for which the alert was raised to the Generative AI service. Upon doing so, this section (see Figure 3) provides insights into whether the deviation in the value of the measure was gradual or sudden or if it was within acceptable limits.

    Figure 3 : The Anomaly Detection section

  4. Diagnosis: If the Would like to enable detailed insights? flag in the AI ASSISTANT SETTINGS page is set to Yes, and if detailed diagnosis is available for the measure for which the alert was raised, then, eG Enterprise automatically uploads the detailed diagnosis data to the Generative AI service. Upon doing so, this section provides additional insights which helps you analyze the reason for the alert raised for the measure at a faster pace. On the other hand, if the Would like to enable detailed insights? flag in the AI ASSISTANT SETTINGS page is set to No or if the detailed diagnosis data is not available, then, this section offers a generic response instead of detailed insights.

    Figure 4 : The Diagnosis section

  5. Best Practices: This section provides best practices and insights on how to avoid the alert in the future.

Figure 5 : The Best Practices section

You can even use the eG AI Assistant even if there are no alerts raised for the measures. For this, you should have checked the check box preceding Normal field against the AI Alert & metric forwarding option in the AI ASSISTANT SETTINGS page. In this case, the eGenie - AI Assistant console will appear as shown in Figure 6.

Figure 6 : The eGenie - AI Assistant console for a normal measure

For a non-problematic measure, the eGenie - AI Assistant console consists of the following sections:

  1. Explanation: This section, as shown in Figure 6, provides a detailed guide explaining what the alert is, the factors contributing to it, its impact, and ways to address the issue. Administrators can use these insights into identify the pain points and suppress the alert within a short span of time.

  2. Impact: This section as shown in Figure 7, offers an analysis of the chosen measure which includes recommended baselines and best practices for improving the performance of the server.

    Figure 7 : The Impact section

Information Exchanged with Generative AI Services

Often, privacy is a concern when integrating eG Enterprise with Generative AI services. Administrators should assert what information is exchanged with Generative AI services and how it will impact their organization, well before starting their integration. To aid administrators make informed decisions, eG Enterprise lists down the information exchanged with the Generative AI services.

If the Would like to enable detailed insights? flag in the AI ASSISTANT SETTINGS page is set to Yes, then, for the problematic measures (i.e., measures for which alerts were raised), the following table lists the section in the eGenie - AI Assistant Console and the information exchanged with the Generative AI services to obtain precise insights into the problem in the section.

Section

 Information Exchanged
Explanation, Analysis Problem description without descriptor and component name/type
Anomaly Detection Test name, Measure name, Measure value, Measure unit, Avg/min/max of that measure, Thresholds and historical data as Base64 Image
Diagnosis Problem description without descriptor and component name/type and detailed diagnosis* (if allowed by the user).
Best Practices Problem description without descriptor and component name/type and metric thresholds

Note:

You can even skip sensitive data in the detailed diagnosis from being passed to the Generative AI services using the below setting:

  • In the <eG_INSTALL_DIR>/manager/config/eg_services.ini file, navigate to the [AI_Integration_Settings] section.

  • By default, the DDSkippedColumns field includes a comma-separated list of IP,PORT,Ip,port,host,server,hostname parameters. This indicates that the detailed diagnosis data with the columns names mentioned against the DDSkippedColumns field will be automatically skipped from being parsed to the Generative AI services. This process improves the security by withholding sensitive data within eG Enterprise.

  • You can even add the name of additional columns from the detailed diagnosis as parameters against the DDSkippedColumns field based on your choice, so that you can avoid these columns from being parsed to the Generative AI services.

For non problematic measures reported by eG Enterprise, the following table lists the section in the eGenie - AI Assistant Console and the information exchanged with the Generative AI services to obtain precise insights into the problem in the section.

Section

 Information Exchanged
Explanation Test name and measure name
Impact Problem description without descriptor and component name/type and metric thresholds