JENNIFER APM

The Statistics menu holds report screens that sum up the performance of a domain and its instances by the day and by the month. Where the real-time dashboards show the state now, statistics show how it went compared with yesterday or last month. CSV export and printing are there so the screens can go straight into regular operations reports.

Both screens work on the first domain the profile's console account can see.

Title bar

While data loads, the search conditions button reads 검색 중... (searching) and, under the mouse, 검색 중단 (stop searching). Clicking it then turns the searching state off, but the search in progress does not stop.

Search conditions window

• Operating hours apply to every daily value except daily visitors. Narrow them to compare business hours only
• Peak condition decides what makes the busiest hour of the day (for monthly, the busiest day)
• Changing the date or the operating hours reloads the target list and checks every target again
• Done (완료) searches again even when nothing changed. With no target checked, it does not search
• Cancel (취소) leaves the date, operating hours and peak condition as edited, and re-checks as many targets as were saved, counting from the top of the list. If you changed targets and cancelled, check the targets again the next time you open the window

Per target conditions window

Choose whether the per instance detail section shows as a table (테이블) or as charts (차트), and check the metrics to show in the Metrics list. The metrics are what the instances report: GC, heap and system memory, CPU, threads, calls, response time, TPS, active services, events, errors, SQL, external calls, fetches, DB connections and more. Calls, response time, TPS, concurrent users, visitors, event count, error count and maximum TPS are checked at first. Apply loads the detail section again and scrolls to it.

Daily System Performance

Summary

The three boxes at the top are domain-wide values.

• Left - the date, domain, operating hours, instances and peak time condition
• Middle - daily visitors, calls, events and errors (all but visitors within the operating hours)
• Right - the peak time (for example 0:00~1:00) with that hour's visitors, calls, response time, TPS, events (domain | instance) and errors

1. Domain performance (Domain 성능)

Six hourly charts: calls, visitors per hour, errors, TPS, response time (ms) and concurrent users. They run from 0 to 24 h, shade the peak time and label the largest value. Hover over a chart for guide lines with the value and time under the mouse.

Under each chart are the day's total (or average) and two comparisons.

• Against the day before (전일 대비)
• Against the same weekday last week (전주 동일 요일 대비)

2. Instance performance summary (Instance 성능 요약)

A row per instance with calls, response time (ms), TPS, concurrent users, events and errors. 2-1. 운영시간 covers the operating hours, 2-2. Peak time the peak hour. Both tables export to CSV.

3. Per instance details, by hour (Instance 별 상세 성능 (시간당))

A table (or charts) per instance, by hour. The columns are the metrics chosen in the per target conditions. As charts, every metric is a one-line small chart.

4. Errors per instance (Instance 별 ERROR 건수)

Error counts by instance and error type: which errors gathered on which instance.

Monthly System Performance

Summary

• Left - the month, domain, operating hours, instances and peak day condition
• Middle - the month's highest daily visitors and highest daily calls within operating hours, each with its date
• Right - the peak day with its daily visitors, calls, response time, TPS, events (domain | instance) and errors

1. Domain performance (Domain 성능)

Choose charts (차트로 보기) or tables (테이블로 보기).

• Charts - calls, daily visitors, errors, TPS, response time (ms) and concurrent users by day from the 1st to the last day, with the peak day shaded. Under each are the total or average and the comparison against the month before (전월 대비)
• Tables - this month's and last month's totals and averages with the change, and a table with a row per day (일별 상세보기, daily details; CSV)

2. Monthly operation status (월 운영 현황)

Last month and this month side by side for every target, with the change on the line below.

Hover over a cell for the unrounded values. The CSV export saves the strings as shown on screen.

3. Instance performance summary (Instance 성능 요약)

Calls, response time, TPS, concurrent users, events and errors per instance on the peak day.

4. Per instance details, by day (Instance 별 상세 성능 (일일))

A table (or charts) per instance, by day. The columns are the metrics chosen in the per target conditions; the monthly screen uses daily visitors instead of visitors per hour.

5. Errors per instance (Instance 별 ERROR 건수)

The month's error counts by instance and error type.

Reading the comparisons

Comparisons with the previous day, week or month show as a percentage with an arrow: ↑ for up, ↓ for down. The percentage is shown as a whole number.

When the value compared against is 0, there is nothing to divide by and the screen shows Infinity% (NaN% in the monthly table view). It means the comparison is with a day or month that had no data; it is not a fault.

CSV and printing

CSV 내보내기 above a table downloads its rows as a CSV file. Numbers are saved as raw values, without the formatting on screen, and a table with no rows downloads no file. CSV export leaves no record in the console.

Print opens the browser's print dialog. If the data has not all arrived yet, it first asks 아직 모든 데이터가 로드되지 않았습니다. 프린트 하시겠습니까? ("Not all data has loaded yet. Print anyway?"). The monthly screen switches section 1 back to charts for printing.

Set as start page

Checking 시작 화면으로 설정 in the title bar's ⋯ makes this statistics screen the first screen the profile's account sees after logging in to the JENNIFER console. The check follows that account's current start page.