Online statistics
Open Journal -> Online statistics and select a project. The section has three pages: Overview, Sample history, and Maintenance. The game server filter selects one server or all servers connected to the project.
Viewing requires the Online and Server Health permission. Clear Server Statistics is a separate permission: without it, the maintenance page and history deletion are unavailable.
Latest samples
On Overview, each server has its own card with actual and published online counts, characters, clans, connection checks, and the last sample time.
| State | Meaning |
|---|---|
| Up to date | A recent sample was received. This does not mean both connections are available: their states are displayed separately. |
| Outdated | No recent sample has arrived. Saved values retain their original time instead of becoming a new zero count. |
| No data | No sample has been received for this server. |
0 | The received sample contains an actual zero. |
- | A value is missing or the selected game build does not provide it, such as the clan count. |
Login and game connection checks are separate:
| Connection | Meaning |
|---|---|
| Available | A connection to the configured address and port was established. |
| Unavailable | The address and port are configured, but the connection failed. |
| Not configured | No address or port is configured for the check. This does not indicate a failed online measurement. |
| Configuration error | Connection settings are incomplete or invalid. Review the server settings. |
A connection check does not replace the online measurement. A successfully measured count remains in cards and charts even when a port is unavailable. If the source read itself fails, no new sample is recorded: the previous sample keeps its original time, or No data remains visible.
Cards may represent different times, so their counts are not added into a combined "online now" total.
Count and period
Actual online comes from the source selected in the game server settings: its database or another online-reading method supported by the game build. Published online includes the server's online display settings and may differ from the actual count. Neither is a count of unique cabinet visitors; use Unique Players for that report.
Select a count, a period, and Refresh. There is no limit on the number of years: you can select several decades when those measurements are still retained. Filter dates, weekdays, hours, and calendar boundaries use UTC; the last selected date is included in full. If that date has not ended yet, calculations stop at the observed time. Individual sample times use the interface's timezone.
| Field | Use |
|---|---|
| Period | Today, the last 7, 30 or 90 days, this year, last year, or custom dates. |
| From / Through | Enter any allowed range, such as January 1 through December 31 of a chosen year. Editing dates switches the preset to a custom period. |
| Group by | Automatic, hours, days, weeks, months, or years. Available in the overview; controls charts and the character/clan peak table. |
The selected count applies to analytical charts and peaks. Sample history retains both values. Moving between pages preserves the selected project, server, period, and grouping.
Annual and multi-year periods
Select This year to view the accumulated samples by month. Automatic grouping uses finer detail near the start of the year; future months are not plotted. This example only has September measurements, so earlier months remain empty.
For several years, set the start and end dates and select Years or Automatic. Keep all project servers or choose one. Automatic grouping considers the period and the number of selected servers; a point can cover multiple years when necessary. Means use the underlying samples, not averages of annual averages. A year without measurements is not treated as zero; changing the period does not restore deleted history.
Last year selects the entire previous calendar year. Without samples for that year, the charts show no data. The latest-sample cards above the charts remain current, regardless of the selected year.
If a long period has too fine a grouping, the page asks you to choose a coarser interval or shorten the period. It does not silently truncate the selected history. For example, select months or years instead of hours for a multi-year report.
Charts and comparisons
- Online by server shows the mean of reliable samples. The previous period has the same elapsed duration. A percentage change is not calculated when the previous mean is absent or zero.
- Recorded samples includes all successful online measurements, including zero. Connection checks counts samples with at least one configured address checked and no invalid connection configuration.
- Successful checks, % is the share of checks where every configured connection was available. An unconfigured connection is not a failure. If there were no configured checks, the percentage is
-. This is not a guarantee of uninterrupted availability between samples. - Peak online shows the maximum for the selected period, all retained history, today, the last seven days, and the current month.
- Average online by weekday and hour compares activity at different hours. It averages samples from the selected servers rather than adding their simultaneous online counts.
- Characters and Clans show changes in game record counts. The table below the charts contains their peaks and reliable sample counts for the selected intervals, not counts of newly acquired or retained players.
Missing chart intervals are not replaced with zero. Automatic grouping uses hours for up to two elapsed days, days up to 90 days, months up to 732 days, and years for longer history. Weeks start on Monday. Months and years use their actual lengths, including February 29. When only part of a month or week is selected, its values include only samples within the selected dates. Cleaning up old history also changes the available peaks and period comparisons.
Sample history
Open Sample history to inspect individual records. The table contains time, server, both online counts, characters, clans, and connection check results.
Select 25, 50, or 100 samples per page. Newest records appear first; Previous and Next open adjacent pages of the current result.
The next page continues the same list. Changing filters restarts navigation. If a history page has expired, refresh the list.
A period without records has its own message and no navigation buttons. This does not mean the online count was zero throughout that period.
History cleanup
Open Maintenance. Select a game server or all project servers, then History type. The status of previously requested cleanup appears below the form.
Enter Before date (UTC) and choose Clear old history. Only records before the start of that date are removed; records on the date remain. Today and future dates are not accepted.
Review the server, history type, and date in the confirmation. Choosing Cancel leaves the data unchanged.
After confirmation, old samples are excluded from this report. In progress shows how many records still need deletion; Complete means no matching records remain. Newer data is preserved.
Repeating the same cleanup does not restore records. An earlier cutoff date cannot undo an existing cleanup either. If a game server is connected to several projects in the team, cleaning its shared history affects all those projects.
Rating snapshots are cleaned independently of Online samples. Check the selected type before confirming: cleaning one type does not delete the other.
Cleanup is irreversible. Removed samples can no longer contribute to charts, peaks, or period comparisons.
Refresh and availability
The page periodically refreshes while visible and while you are not editing filters. Refreshing the page does not trigger a new game server measurement: between measurements, it shows the latest saved values and their time. A warning appears if recent samples are still arriving and totals may change.
A loading failure produces an unavailable message, not an empty successful report. Refresh again after connectivity recovers. A project with no connected servers and an account with no available projects each receive an explicit empty state.