Dashboards

Every Project opens on a built-in default dashboard. Everyone sees the same default, and nobody can change it. Each member can also keep several named personal dashboards per Project, for example “Traffic”, “API health” and “Release watch”. Changing yours never changes anyone else’s.

The default dashboard

The default dashboard is defined by OpenRUM and always available at the top of the dashboard list. It reads top to bottom:

• Headline numbers: Sessions, Page Views, users, error rate, API failure rate, and LCP, INP and CLS at P75, each with its change against the previous period.
• Trends: Sessions against the previous period, Page Views and users, Page Views by device, API request outcomes, errors and LCP.
• What to look at first: top pages, Page Views by browser and country, errors by release, a country overview table, top Issues and the slowest APIs.

The default uses every kind of module at least once, so it also shows what a dashboard can draw.

You can edit the default like any other dashboard. When you save, OpenRUM creates a new personal dashboard with your changes and opens it. The default stays exactly as it was, for you and for everyone else.

Personal dashboards

Open the dashboard name at the top of the page to switch between the default and your own dashboards, create one, rename it, reorder the list or delete the current one.

• A new dashboard can start blank, from the default layout, from a template (e-commerce overview, traffic and Sessions, performance, API health, error analysis, business metrics) or as a copy of the current one. The e-commerce overview shows revenue, orders, average order value and the purchase funnel; it expects view_item, add_to_cart, begin_checkout and purchase Custom Events, with the order total in a purchase measurement named amount.
• Each dashboard keeps its own draft and saves independently. Saving one that changed on another device is refused rather than overwritten; reload it and edit again.
• The plain dashboard address opens the dashboard you last used on this device. Each dashboard has its own address, and links keep the selected time and Environment.
• Each person can hold up to 20 personal dashboards per Project, each with up to 24 modules. The default does not count toward the limit.

Modules

Module	Shows
Stat	One value for the selected range, with the change against the previous period
Time series	One or more metrics over time as area, line or bar, optionally with the previous period as a dashed line
Stacked time series	Parts of a whole over time, either several disjoint metrics or one metric split by a dimension
Breakdown	One metric by dimension as ranked bars, a table or a donut
Ranked table	One metric by group with its change and a small trend per row
Metric table	Several metrics side by side per group, sortable by any column

The module library is organized by data domain: traffic and Sessions, performance, API, errors, business metrics and user behavior. Blank modules for building your own are under Custom. Hover the gear at the top of the page to edit the dashboard or add a module.

Modules follow the page’s time range and Environment. A module can also carry its own filters, such as release, route, country, browser, device, event kind or event name, depending on the metric.

The metric catalog

Modules choose metrics from a catalog defined by OpenRUM, not from a query editor. Every metric reads an existing aggregate, so dashboards stay fast and the catalog decides which combinations are meaningful.

Group	Metrics
Traffic and Sessions	Page Views, users, Sessions, errors, API Requests, error rate, API failure rate
Web Vitals	LCP, INP and CLS at P75
API	Requests, failures, 4xx/5xx/network outcomes and their rates, latency P50/P75/P95
Errors	Issue events, affected users and Sessions, active Issues, new Issues
Business metrics	Sum, sample-rate-weighted sum, average, minimum, maximum, P50 and P90 of a Custom Event measurement
User behavior	Event count, users and Sessions for page views, route changes, clicks and Custom Events

Compare several events on one chart

To see a payment flow, add a time series, choose Event count and split it by Event name. Then pick the events to compare, for example checkout_started, payment_succeeded and payment_failed. Each event becomes its own line on one axis, in the order you picked them, up to nine lines.

• Without picked events, the chart shows the most frequent events instead. Filter by event kind to keep page views and clicks out of it.
• An event that did not happen in a bucket leaves a gap rather than a zero.
• Event counts may be stacked, because each event is counted once. Users and Sessions may not: one person can trigger several of the events.
• Success plus failure does not have to equal the number of starts. Some people abandon a flow, so a line chart is usually the honest choice.

Rules that keep the numbers honest

• One unit per chart. Metrics drawn together share a unit. Counts scaled back up by the sample rate cannot share an axis with counts that were not.
• Only parts of a whole stack. A stacked chart needs metrics that add up, such as the three API outcome rates, which share one denominator. Error rates, percentiles and distinct users do not add up across groups, so they have no “Other” group, no shares and no donut.
• A 4xx is not a failure. API failures are network-level failures and 5xx responses. The outcome chart shows 4xx separately, as a share of all requests.
• New Issues are Issues that first appear in the range, or return after 30 days without an occurrence. “First seen ever” cannot be answered in a bounded query and is limited by aggregate retention anyway.
• Error metrics are drawn in five-minute buckets or wider, because that is how they are stored.
• Web Vitals are P75 only on dashboards. Other percentiles, FCP and TTFB are on the Performance page.
• Business metrics need a measurement key, such as amount, and can be split by country, device, browser, source or a Custom Event attribute. They cannot also be filtered by country or browser.
• Behavior metrics split by one dimension at a time. Event name, country, device, browser, source or a Custom Event attribute can each split a chart, but a chart split by country cannot also be filtered by browser.

Every chart follows the shared point budget of about 30 buckets. A bucket with no data is left empty rather than drawn as zero.

When a module shows an error

A module whose query would scan too much reports that its query exceeded the budget. Shorten the time range or add an Environment, release or route filter. Other modules keep working.