Ten reports, five export formats, scheduled email delivery that carries the figures in the body, seventeen dashboard widgets, and a timesheet that shows you the empty days. Drilldown from any figure, row, cell or chart segment straight to the tickets behind it.
Reporting is where most trackers hand you off to a paid add-on. Here is what is in the product.
A report earns its place by answering a question someone actually asks in a review meeting.
Ad-hoc reports from a TQL query or a natural-language AI prompt. Breakdowns by status, priority, assignee and component, plus a created-versus-closed trend.
Optional AI analysis suggests charts, and a conversational refine query loop lets you narrow the question without starting again. Chart images rendered in the browser can be posted back so the generated PDF carries the same visuals you were looking at.
Tracked time per ticket plus the work-log detail lines behind it, costed at the project's blended day rate in the organisation's currency. The detail lines are the point — an invoice line nobody can explain becomes an invoice line somebody disputes.
Allocated versus spent over time, bucketed by day, week, month or year, derived from logged minutes multiplied by the day rate on the convention that eight hours is one man-day. A billable-only toggle, a consumption bar, and cumulative and man-day columns.
Compliance rates, the breach list, response-time analytics, a breach heatmap and per-project compliance. See SLA for how the timers themselves work.
Mean time to resolution and its trend, fastest and slowest resolutions, first-contact-resolution rate overall and per project, the reopen trend, and a running backlog trend.
Reopens and first-contact resolution sit next to MTTR on purpose: a team can always improve MTTR by closing things early, and the reopen trend is what catches it.
Per-status and per-ticket time rebuilt by replaying status-change history rather than by reading a stored counter, along with average age, average number of transitions and the worst status.
One honest detail: the currently-open segment is reported but marked not counted when the ticket's present status is DONE or CANCELLED, so a closed ticket does not accumulate phantom time in its final state.
Workload: open, high priority, overdue, due soon, on hold, oldest, and the unassigned load. Delivery: resolved versus filed, average and median time to resolve, reopen rate, estimate accuracy and how much is unestimated. Effort: logged time, billable time, days worked and utilisation against your own working day, sliced per person, per project or per bucket.
Three lenses, kept apart. There is no composite productivity score, deliberately.
Per-ticket tracked time with a work-log drill-down, so a total on a ticket resolves into the individual entries and the people who made them.
Usage and cost dashboards at organisation and platform level, with an inline monthly budget editor and a cost-trend chart. See AI for the provider list, including local models that cost nothing to run.
The /stats surface: totals and status splits, age distribution, overdue analysis, assignee performance, bottleneck analysis, priority aging and workflow health. Detailed in the next section.
Backed by materialised views with a scheduled refresh — the default is every five minutes — plus a Spring cache, so the page is fast on a large install.
Statistics, generator, budget, resolution, team, time-in-status, ticket-time, timesheets, team timesheets, SLA, AI usage, audit logs and single tickets all export.
excel and xls both resolve to XLSX
SVG charts export as .svg or .png
Every report page shares one export menu and a print stylesheet
Saved TQL reports and six metric report kinds — GENERATOR, BILLING, BUDGET, SLA, RESOLUTION and TEAM.
nextRunAt advances before dispatch, not after. A crash mid-send costs you one report; the alternative costs you the same report four times and a conversation about trust. History is retained per schedule and pruned.The organisation-level policy sits above all of it: schedules can be enabled or disabled organisation-wide, external recipients can be allowed or forbidden outright, and a maximum recipient count applies — 20 by default, with a hard ceiling of 50. Administrators get an organisation-wide view of every schedule that exists and can disable any of them, plus per-schedule delivery history and a run now for when someone asks for the numbers before the cron does.
A report you cannot interrogate is a report nobody trusts.
Every report page shares the same export menu, print stylesheet, schedule-this-report dialog and drilldown. Click a figure, a row, a cell or a segment of a chart and Trakr opens the tickets behind it.
It is a property of the report surface rather than a feature bolted onto one chart, which means the answer to "where does that 47 come from?" is always one click, on every report.
Per-user layouts with tabs you can create, rename, drag to reorder, give an icon and delete, over a widget grid.
The seventeen widget types: My Tickets, Created by Me, Projects Overview, Recent Activity, Saved Filter, Watched With Updates, AI Search, Recent Items, My Open Bugs, Due This Week, Unassigned Critical, Past Due Date, AI Assistant, Quick Add Ticket, Timesheet, Links (Metro-style tiles, personal or organisation scope) and My Requests.
Share a layout snapshot with named colleagues, browse what you have shared and what has been shared with you, and apply someone else's layout into your own.
The rule that makes this safe: widget data is always rendered with the viewer's identity. Applying a manager's dashboard gives you their arrangement, not their access. A shared layout can never become an accidental data leak.
A saved report is a user-owned definition carrying a name, a TQL query, an AI prompt, chart configurations and any AI analysis. It can be kept private, shared organisation-wide, or marked as a favourite, and edits are owner-gated.
Ticket-table widgets get the full column manager — sortable columns, a column chooser, grouping and CSV export — the same table behaviour used on every list surface in the product.
One storage model, no parallel time-entry type, and a grid that includes the days nobody logged anything.
Work logs carry time, a work date, a description and a billable flag. There is no separate time-entry type sitting alongside them, which is why every time figure in the product — Billing, Budget, Ticket time, the Effort lens, the timesheet — reconciles to the same rows.
Time is entered in notation: 30m, 2h, 1d where a day is eight hours. Editing and deleting are gated to the author or an administrator. Totals are available per ticket and per author over a date range.
The active timer starts, stops and discards per ticket, survives a page reload, and opens the log-work dialog when you stop it.
A manager can log on behalf of another person, from the ticket or from the team timesheet view — because the alternative is a manager typing someone else's hours into a spreadsheet, which is the same act with none of the audit trail.
A week or month day grid with a per-day meter against the organisation's day limit, a hide-empty-days toggle, and a per-day pane of entries. Click a day or an entry to log or edit, with a ticket search picker, description, time and billable flag.
The period bar offers Today, this and last week, this and last month, and a custom range bounded to a year. Every day in the period is present — the gaps are the information, and a list that quietly omits them is not a timesheet.
Buckets by day, week, month, quarter or year — auto-picked from the range — grouped by period or by person, with two day metrics shown together: "days logged", meaning distinct dates, which becomes person-days across a team, and "FTE days", meaning minutes divided by the organisation's working day.
They answer different questions. Days logged tells you how many days somebody touched the system; FTE days tells you how much work that came to. Showing only one of them is how time reporting starts producing numbers people argue with.
Period-over-period comparison runs against the previous period, the previous year or year-over-year, spanning up to the last three years, with charts and project totals.
View one person's timesheet or several at once, with a roster and group support, and log time on behalf of another person from the same screen.
Organisation-boundary enforcement has a deliberate shape: a subject outside your organisation returns 404, not 403. A 403 confirms that the person exists; a 404 tells a prober nothing.
Alongside it, timesheet reports run per project, grouping time by person, and per user, grouping time by project — each with a summary and a detail view.
Two fields — regularMinutesPerDay and overtimeMinutesPerDay — each with three states: inherit (null), unlimited (-1), or a number. Resolution climbs organisation → tenant → platform → built-in, independently for each field, and the effective view reports which scope produced each answer.
It ships unlimited, so nothing is capped until you decide to cap it. The same figure drives Team Performance utilisation and the timesheet's FTE days, so there is one definition of a working day rather than two that drift apart.
CSV, XLSX, PDF, JSON and XML, plus print. Beyond the standard set there is a multi-month PDF — pick the months, take a whole-year shortcut, step by year — with weekend shading and month separator pages, and a several-people by several-months PDF for a team.
Statistics export in the same formats, including the compared-period variant, so the comparison you are looking at is the comparison that leaves in the file.
There is no timesheet approval workflow. Approval machinery — requests, reviewers, states — belongs to change management, and duplicating it in time entry would give you a second, weaker approval engine to maintain.
There is no composite productivity score. Team performance deliberately keeps Workload, Delivery and Effort as three separate lenses. The moment those collapse into a single number, that number becomes the target — tickets get split to raise throughput, estimates get padded to protect accuracy, and the metric stops describing the work it was meant to measure. Three numbers that mean something beat one number that gets managed.
Yes. Every report page shares the same export menu, print stylesheet, schedule-this-report dialog and drilldown: click any figure, row, cell or chart segment and Trakr opens the tickets behind it. A report you cannot interrogate is a report nobody trusts, so drilldown is a property of the report surface rather than a feature of one particular chart.
Exactly five: PDF, CSV, JSON, XML and XLSX, with excel and xls treated as aliases for XLSX. An unknown format is refused with a 400 that names the supported list — it is never silently substituted for something else. PDF is drawn server-side with OpenPDF including proportional bars, and XLSX is written with Apache POI as one sheet per section carrying real numbers rather than strings.
Yes. Saved TQL reports and six metric report kinds — GENERATOR, BILLING, BUDGET, SLA, RESOLUTION and TEAM — can be scheduled daily, weekly on an ISO day, or monthly on a day-of-month that is clamped for short months, each at a chosen hour. Seven reporting windows are available: last 7, 30 or 90 days, last 6 or 12 months, this month, or the previous month. Recipients can be internal users and free-text external addresses, and they are re-resolved on every run.
No, deliberately. The mail carries the figures in the body rather than as an attachment, so the numbers are readable on a phone without opening anything. A scheduled run also never calls AI — a nightly job is not the place to spend tokens or to introduce non-determinism into a figure someone will act on.
Every day in the period is present in the grid, including empty ones, with a hide-empty-days toggle if you want them out of the way. Each day carries a meter against the organisation's day limit and a pane of its entries. This is the difference between a timesheet and a list of work logs: the gaps are the information.
Neither, and both absences are deliberate. Approval machinery belongs to change management, not to time entry. And team reporting exposes Workload, Delivery and Effort as three separate lenses with no composite productivity score, because collapsing them into one number invites the number to be managed instead of the work.
Ten reports, five export formats, scheduled delivery and a real timesheet — included at every tier, on every deployment.