Sum hours, story points or any number field. Without rebuilding the report every time.

A total that hides what is missing from it is a total nobody can defend. So the grid counts the issues carrying no figure at all, cell by cell: Mark what is missing corners every figure that left out 20% or more of its issues, and hovering over a figure says how many it left out and that they add nothing to it.

The grid also says when a read was cut short, and why. An axis with more values than a table can hold gathers the rest into one row called Everything else, where they still count, rather than dropping them. And where an issue genuinely belongs in three rows, it says how far that pushes the margins past the grand total instead of quietly picking one. The whole arithmetic is public.

A gadget gets the width its dashboard's layout gives it, and in a two- or three-column layout that is 346 pixels. Ten statuses across the top want far more.

A column that narrow is where this product lives rather than an edge case, so what happens there is decided rather than suffered: the table scrolls sideways with the row labels and the Total column held still, it says underneath how many columns are off the screen and that their figures are in the totals, and where the other arrangement would fit it offers the swap by name. Drag the frame on the demo to see it at that width, or read what it does in a column.

• Crosstab draws the grid.
• Crosstab Total draws one figure and the number of issues behind it.
• Crosstab Change lists what has moved since a grid somebody kept or published, largest first.
• Crosstab Sprints draws the last sprints of a filter: what was in each at the start, what was added after, and what was closed by the end.
• Crosstab Chart draws the grid as bars stacked or side by side, as each row's shares of a hundred, or as a line for each of its six largest rows. Its bars show the work still open: a column that is a status in Jira's done category, such as Done or Canceled, is left off and named underneath, unless finished work is all there is to draw. Up to seven columns are drawn one by one; past seven, the six with the most in them are, and the rest are drawn together. The table under the chart holds what the chart draws; every column is in the grid Crosstab draws.

A gadget can take its settings from another Crosstab gadget on the same dashboard: Use these in its settings links the two, and a change made to one shows up in the other.

Gadgets with the same settings on one dashboard read Jira once between them, each reader with their own permissions. Crosstab Sprints is the exception: it reads the history of every issue in the filter that has been in a sprint, and shares that read with no other gadget.

Open at page width opens the grid across a whole page, with all nine views and the comparisons a dashboard column has no room for.

Jira's Two Dimensional Filter Statistics gadget answers how many. This one answers how much: hours, story points, an estimate, or any numeric field in your site, across any two axes it is allowed to draw. The one number it will not add up is Work Ratio, which Jira works out per issue as a percentage, and a sum of percentages is not one.

Sprint scope is an axis that splits every sprint into the work that was in it at the start, the work added after the start, and the work the record cannot place (Start not in the record), read out of each issue's own history. Put it beside sprint and each row is a sprint's plan against what arrived later, in points or hours or any field you total. Where the start is drawn, and why.

A site of ten people or fewer pays nothing, and one of twenty-five pays $125 a year for the whole site, however many dashboards carry it.

A field summed across two axes is something a CSV and a spreadsheet will give you once; what they will not do is go on giving it to you, on the dashboard, next week.

Three things here are the reason to pay for Crosstab rather than export: a grid that says what each figure left out; Keep, which stores the grid as it stands, so that Crosstab Change can say how far every cell has moved since; and the Sprint scope axis above. All three are on the demo.

Already have a pivot gadget, from a bundle or a free one? It sums a field across two axes too. Ask it the same three things: does it say what each figure left out, and when a read was cut short; can it keep a grid and say how far each cell has moved since; and can the same grid split a sprint into what was in it at the start and what was added after?

A kept grid is the keeper's own: a colleague opening the same gadget sees the grids they kept, and Publish this grid is how one reaches everybody else.

Under about three thousand issues, a read finishes and every figure here is a complete total. Between three and twenty thousand one read does not finish on most sites: the gadget has its resolver's clock rather than your patience, and where it stops early the grid says so on its face, marks the figures it could not finish, leaves nothing out silently, and offers Finish this read, which reads on from where it stopped and adds to the same figures.

One grid reads at most twenty thousand issues, however many times the read is continued, and stops there by design, so a filter larger than that is one to narrow before it is one to draw. That reach is arithmetic on a page of a hundred issues, which makes it an estimate of your own site rather than a property of this product — both the sum and what a grid looks like when a read is cut short.

However large the site, the gadgets people open on it spend at most 13,000 points of Jira's hourly allowance between them, the same for a site of ten people and of a thousand: about four first opens in an hour of a three-thousand-issue grid that adds up a field. A grid held for five minutes, and a kept one, ask Jira nothing. What each open costs, and how to work it out for your site.

Free for ten users or fewer. Past that, from $0.50 per licensed user a month on your Jira site, on a graduated scale that gets cheaper as the site grows. Thirty days to try it first.

Every band, and what a year comes to →

Jira is read on a schedule only where somebody switched one on, and there is no external host: the built bundle reaches nothing but your own site, and the manifest Atlassian shows you before installing declares no host at all. What the app keeps goes into its storage on Atlassian, and the vendor sees only counts and short log lines, read through Atlassian's developer console, which your administrator can switch off — the privacy page lists each of them. No axis can be a person — not assignee, not reporter, not any user field — because a total broken down by a person is a metric about that person.

• Live demothe gadgets themselves, on an invented board, with nothing to install
• Getting startedhow to configure it, and how to read what it draws
• Methodthe whole arithmetic behind every figure in the grid
• Securityhow the app is built and operated, and how to report a flaw
• Privacywhat is read, what is stored, and why none of it leaves Atlassian
• Supporta question, a fault, or something that looks wrong, and when you will hear back
• Pricingwhat it costs, what a year comes to, and who pays for what

Crosstab is on the Atlassian Marketplace — install it there, or try the gadgets here first.