Getting Started
This page follows a single pass: opening the console screens for the first time, picking odd transactions on a dashboard, and reading the profile of one of them. Each screen is covered in detail in its own page.
Before you start, add a JENNIFER Console connect profile as described under console setup in the Install Guide. A JENNIFER APM profile for transaction collection alone does not open the screens.
1. Open the System Admin dashboard
In the left menu, open Dashboard > System Admin under JENNIFER APM.
- With one
JENNIFER Consoleprofile, the dashboard opens straight away - With two or more, a list of profile names and endpoints comes first. Click the console you want and the dashboard opens
The profile you picked stays at the end of the address as ?profile=. Other menus open with the same profile, so to look at another console, open the menu again and pick, or change the profile in the address.
2. Read the dashboard
The dashboard is laid out as follows.
- Header - the name of the domain group in view is on the left; the screen name is in the middle
- Domain bar - right below the header, the domain's instance names and the instance counts by state
- Chart cards - cards such as 액티브서비스 (active services), TPS, 동시사용자 (concurrent users), 응답시간 (response time), 시간당 호출 건수 (calls per hour), X-View and EVENT refresh every few seconds
The starting point is spotting what looks different from usual: a spike in the response time chart, active services piling up in the long-running band (8 seconds or more, red), or a warning on the event card. What each card means is in Dashboards.
3. Pick transactions on X-View
The X-View card plots each finished transaction as one point. The horizontal axis is when it finished and the vertical axis is its response time, so points floating high are slow transactions.
- Look for points crowded near the top, or reddish points (transactions that ended in an error)
- Drag a rectangle around them with the mouse
- When you release the mouse, the X-View 트랜잭션 (X-View transactions) window opens with the transactions inside the rectangle in a list
Clicking the X-View chart once and pressing Ctrl+B is the same as selecting the whole chart. Use it to open every transaction currently shown at once.
4. Narrow down the list
The X-View transactions window has filters on the left and the list on the right.
- Filters on the left - counts by response time range, transaction kind (ERROR, asynchronous and so on), application, Domain, Instance, client IP and HTTP status. Checking one narrows the list to it
- 트랜잭션 데이터 (transaction data) tab - per transaction, the application, instance, start and end time, response time, and the method, SQL, External Call and CPU time within it as bars and ratios
- 애플리케이션 맵 (application map) tab - groups the selected transactions by application to show which application spent the most time
Rows with a long response time and a large share of SQL or External Call time are the ones to open next. The filters and tabs are covered in X-View Transactions.
5. Read the transaction detail
Click a row in the list and the transaction's detail opens in the same window.
- Top - the application name and the response time, CPU, Method, SQL and External Call times
- Left - the time spent in the Frontend and server parts, the share of Method, SQL and External Call in the server time, and the socket, file, message and ERROR lists
- Tabs on the right - 타임라인 (timeline: when each call started and how long it took), Flame 그래프 (flame graph), 콜 트리 (call tree: how calls nest), 텍스트 (text: the full profile), MSA (calls to linked services)
Find the longest bar in the timeline, and check in the call tree which SQL or external call it is. That is the answer to "where was it slow" for one transaction.
To go back to the list, press the arrow at the top left or press Esc.
6. Widen the view
Whether what you found in one transaction is a one-off or a pattern is checked over a wider scope.
| What to check | Where |
|---|---|
| Transactions of the same application in an earlier period | Analysis > X-View - Analysis |
| What code was running in that period | Analysis > Stacktrace Flight Recorder (SFR) - Analysis |
| A day's or a month's metrics and errors | Statistics > Daily System Performance, Monthly System Performance - Statistics |
| Anomalies against a baseline, comparing windows | Query Commands |
| All of the above from one question | AI Agent Root Cause Analysis |
Next, Dashboards explains how to read each card.