Dashboard Patterns

A number means nothing until the viewer knows which way it moved and by how much.

A person needs to know the current value of one number, and whether it's fine, before they look at anything else.

An average hides the slow requests users actually feel; the viewer needs the tail.

The viewer needs to know which few items dominate, not the distribution of all of them.

A raw count misleads; the viewer needs it normalized by a denominator they understand.

The viewer needs the shape of the recent past next to the number, without the cost of a full chart.

The number only matters relative to a goal, and the viewer needs to see distance to it.

A change in the line has a cause, and the viewer shouldn't have to go find it elsewhere.

The viewer needs to know where a panel's data comes from and whether that source is healthy.

Two people read the same tile and mean different things by it.

The viewer is about to act on a number and doesn't know how old it is.

Too many series on one chart; the viewer needs to isolate the ones that matter without leaving the chart.

Every question about a metric starts with 'over what window', and the control has to be fast for the common windows.

A metric is moving, and the person watching it can't tell whether the movement is normal for this hour on this weekday or the start of an incident.

Clinical staff need every patient's state and location in one board, ordered by urgency.

Daily activity over a year, read for streaks and gaps.

Retention differs by when people arrived; the viewer needs cohorts side by side over time.

The viewer needs exact values for many items and wants to sort and scan them.

People move through steps and the viewer needs to see where they leave.

One value against a range, read at a glance. Few and NN/g both argue against it; this entry says when it survives that critique.

The items have locations and their state is best read spatially.

Thousands of events per second can't be drawn as points; the viewer needs density.

The viewer needs the shape of a population, not a summary statistic of it.

Hundreds of hosts; the viewer needs to spot the hot ones without a table of hundreds of rows.

Something is happening now and the viewer needs to watch the raw events as they arrive.

A trader needs to see supply and demand at every price right now.

An operator needs live values drawn on the physical layout of the plant.

Quantity flows from sources to destinations and the viewer needs to see the split.

The viewer needs to know what depends on what, and which edge is unhealthy.

Many series would tangle on one chart; the viewer needs them on identical axes side by side.

A total is made of parts, and the viewer needs both the total and the mix over time.

The viewer needs to see when something was up, down, or degraded, over days, in one row.

A flat metric has a line it must not cross, and the chart should show where that line is.

The viewer needs to see how a value changed over time and spot the moment it changed.

One request touched twenty services and the viewer needs to see where the time went.

Today's line means nothing without last week's line under it.

Selecting something in one panel should narrow every other panel to it.

The overview shows that something is wrong; the viewer needs to get to what, in one click.

The chart shows shape; the viewer needs the exact value at one moment without leaving it.

The team re-selects the same five filters every morning.

The dashboard has sixty panels and the viewer knows the name of the one they want.

One dashboard layout is right for fifty services, forty hosts, or three environments, and nobody should build it fifty times.

The interesting thing is ten minutes wide on a seven-day chart.

The dashboard grew to sixty panels and the viewer needs the top twelve most of the time.

A small panel needs to become a big one for a moment.

The on-call engineer is looking at this on a phone at 3 a.m.

One page can't hold it and the sections have different audiences.

Many independent panels need to share one screen without fighting for it.

The dashboard is on a wall, ten feet away, with no mouse.

The viewer needs to start wide and end narrow without losing their place.

Panels belong together by meaning, and the layout should say so.

Navigation between many dashboards needs to be persistent without stealing the canvas.

The same data serves different questions and each deserves its own page.

An expert wants everything on one screen and will trade legibility for it.

The first row of the page has to answer 'is it okay' before the rest asks 'why'.

Different viewers care about different numbers and want to arrange them.

A non-expert needs one obvious reading order and no configuration.

The viewer works through a queue and needs the list and the item at once.

Twelve services on one chart need twelve colors a person can tell apart.

Ops dashboards live on dark screens all night and the chart colors have to survive it.

If everything is colored, nothing is; normal should be gray.

Finance readers expect up-green and down-red, and a fifth of them can't tell the two apart.

State has to be readable before the number is.

Magnitude and direction need encoding in color, and the wrong ramp lies.

The chart shows the line to watch; the viewer wants to be told when it's crossed.

The viewer didn't look at the right chart, and the product should tell them what it noticed.

Two people are looking at the same spike and need to talk about it in place.

The person who needs the dashboard has to be able to make it.

There's nothing to show, and the page has to say why and what to do.

One panel's query failed and the viewer must not mistake it for a zero.

The viewer needs to narrow a large set by several attributes and see what's applied.

Forty panels are querying and the page shouldn't look broken for two seconds.

The viewer wants the dashboard to know what 'good' is.

The dashboard has to leave the tool and keep meaning something.