WhaTap AI usage guide

This document guides you, by scenario, on how to use WhaTap AI in real operational situations. For an overview of WhaTap AI features, refer to WhaTap AI introduction; for the first-time flow, refer to Getting started; and for the screen layout, refer to UI guide.

How to ask effectively

WhaTap AI understands natural language questions as they are, but the more specific your question, the more accurate the answer you receive. Including the following three elements in your question helps.

• Target. Specify which metric or resource you want to view. For example, specify response time, CPU usage, or a particular transaction.

• Period. Include the time range to query, such as "the past hour" or "this morning." If you do not specify a period, it follows the time range of the screen you are currently viewing.

• Condition. Adding the condition you want, such as "slowest," "exceeding 90%," or "with errors," lets you narrow down the results.

Tip

WhaTap AI automatically recognizes the screen and widgets you currently have open and the time range you selected. You can keep asking based on the context you are viewing, without describing conditions such as "on this screen" or "with the current time range" every time.

Usage by scenario

Note

The responses below are examples to aid understanding. Actual responses vary depending on the screen you are viewing, the time range you selected, and the collected data.

Quickly diagnose performance anomalies

When you suspect performance degradation, ask directly in natural language instead of moving between screens to find metrics. Based on the metrics shown on the screen, WhaTap AI analyzes trends, anomalies, and even improvement directions.

"Show me the transactions with the slowest response times in the past hour"

It organizes the transactions with slow response times based on the past hour and shows a response time trend chart of the slowest transaction. The answer also displays the name, unit, and a brief description of the metrics used.

You can also ask like this.

• "Which service had the most errors this morning"

• "Analyze the point where CPU usage spiked on the current screen"

Build the dashboard you want

Without choosing widget types and metrics yourself, just state your analysis purpose and WhaTap AI picks charts and attaches them to the response, and when needed, assembles a dashboard layout that groups multiple widgets. The output can be saved and reused as is.

"Build a chart that shows response time and TPS together for the last 24 hours"

It picks a chart that shows response time and TPS together over the last 24 hours and attaches it to the response. If your intent is to view multiple metrics on one screen, it assembles a dashboard layout that groups the widgets, and the output can be saved as is.

You can also ask like this.

• "Gather this service's CPU, memory, and network usage on one screen"

View multiple projects on one screen

When you want to check the status of multiple projects or multiple products together, WhaTap AI automatically creates an integrated Flex Board. You can grasp the overall operational status in one place without opening each project's dashboard one by one.

"Compare the response times of all projects in operation on one screen"

It creates an integrated Flex Board that gathers the response times of multiple projects in one place. You can compare the overall status side by side without opening each project's dashboard separately.

Configure alert rules

WhaTap AI suggests an appropriate rule and settings for an alert request and guides you to the relevant event configuration screen. You can apply the suggested rule directly from the response, and if the applied result differs from your intent, you can revert the most recent application and restore the previous state.

"I want a Critical alert when server disk usage exceeds 90%"

It suggests a Critical alert rule and settings with disk usage exceeding 90% as the condition, and guides you to the relevant event configuration screen. You can apply the suggested rule directly from the response, or revert to the previous state after applying it.

You can also ask like this.

• "Send an alert when a transaction with a response time over 3 seconds occurs"

Check the meaning and unit of a metric

When the unit or meaning of a number shown on a dashboard is unclear, you can ask about it directly. WhaTap AI provides the name, unit, and type of the metric used in the answer along with a brief description.

"What does Working Set memory mean"

It explains the meaning, unit, and type of the metric with a brief description. The metric used in the answer is shown with its name and unit, so you do not have to guess the meaning from numbers alone.

You can also ask like this.

• "What unit are the values in this chart"

Analyze screens and events

When you enter the target you want to analyze in the input box at the top of the console, WhaTap AI analyzes the cause and suggests next steps based on the context of the screen you are currently viewing and the events that occurred.

"Tell me the cause of the event that just occurred"

It recognizes the context of the screen you are currently viewing and the events that occurred, and analyzes the cause and suggests next steps. You can continue receiving analysis based on what you are viewing, without closing the screen and re-explaining the conditions.

You can also ask like this.

• "Analyze the current screen"

Using the response

In addition to the text answer, a response may include diagnostic cards, follow-up question chips, charts, dashboards, and alert rule apply buttons.

• Diagnostic summary card: WhaTap AI organizes the analysis result as a card. It shows the key conclusion and severity (Critical / Warning / Info) using color.
• Follow-up question chip: Additional questions worth asking are suggested as chips below the response. Clicking a chip sends the corresponding question directly.
• Chart: Charts attached to the response can be expanded for viewing or saved to a dashboard as is.
• Dashboard layout: When WhaTap AI suggests a layout that groups multiple widgets, you can save it as is and reuse it.
• Integrated Flex Board: When you want to view the status of multiple projects on one screen, WhaTap AI suggests an integrated Flex Board. Saving it lets you use it as an entry screen for operational status checks.
• Alert rule apply: When an alert rule apply button appears in the response, review the rule content and apply it with a single click.

Usage tips

• Keep asking based on the screen you are viewing. Without closing the screen and re-explaining the conditions, you can continue your questions based on the open screen and time range.

• Use it in multiple languages. It recognizes the input language among Korean, English, and Japanese and responds in the same language. A multilingual team can use it together, each in their preferred language.

• Send feedback on responses. You can send feedback with the Like or Dislike button on a response. Selecting Dislike lets you also choose a reason such as accuracy, relevance, or length, and the feedback you provide is used to improve response quality.

Related documents

• WhaTap AI introduction. Check the value and key features of WhaTap AI.

• Getting started. Check the flow and prerequisites for using WhaTap AI for the first time.

• UI guide. Check each area of the WhaTap AI screen and how to use it.