Skip to content

v2026.07.21.1

Choose a tag to compare

@jordibrouwer jordibrouwer released this 22 Jul 22:08
· 1515 commits to main since this release

Monitoring statistics can be opened at full size — a monitored row in the health view had room for a 24-hour uptime figure and one ping, while the report behind it already carried 7-day and 30-day windows, the check count, and the outage list. A button on the row strip now opens all of it in a large view, without leaving the list.

Health view

  • new Enlarge a monitored row's statistics — the button at the end of the monitor strip, or i on the selected row, opens a modal with a large response-time chart (min / average / max marked, tooltip per point), uptime for 24h / 7d / 30d side by side with the sample count behind each, a taller heartbeat bar, interval and last-check time, and the full outage list. It reads the report already on screen rather than fetching again, so it opens instantly and shows exactly what the row below it shows. A window with no samples reads no data instead of 0%, so a newly enabled monitor is not misread as a day of downtime. The button is absent on a monitored bookmark that has not been checked yet — there is nothing to enlarge (dashboard-health.js, health-view.css, locales/{en,nl,de,fr}.json).
  • new i in the row shortcuts — added to the keyboard legend under the feed alongside s, p, c and m. On a row with no monitoring data the key falls through untouched rather than opening an empty panel.
  • new Documented under Config → HelpHealth & status gains a paragraph on the enlarged view: what it shows, that it reads the report already on screen rather than fetching, why an empty window reads no data instead of 0%, and when the button is absent. A matching entry joins Tips & tricks under Keeping it healthy. Both are picked up by the help filter, which matches on section text (locales/{en,nl,de,fr}.json, config-help-tips.js).

Fixes

  • fix Outage lengths always showed as "0s" — the incident list read duration from each outage, but the server sends the field as durationMs (HealthIncident.Duration), so every closed outage was formatted from undefined and rendered as 0s. A twelve-minute outage now reads 12m. This was easy to miss while outages were only visible inside the collapsed score panel; the enlarged view puts them in plain sight (dashboard-health.js).