Dashboard

This page is also in the Groundskeeper wiki: https://github.com/rsjsoftwareltd/groundskeeper/wiki/Dashboard

The Dashboard is the page you land on after signing in. It is the working screen for whoever looks after IT: an activity feed on the left, and the state of every connector on the right.

Open it at https://<server-address>:8443, or from the Dashboard item in the sidebar. Admin and technician accounts see it. An SLT account does not -- it gets the SLT Dashboard and Cyber Essentials Readiness only.

The page refreshes itself. The feed and the connector list update every minute, and the AI summary every five. You do not need to reload it.

The activity feed #

The feed lists problems, not good news. Connectors only add an event when they find something wrong, or when something that was wrong has recovered.

A quiet feed means a healthy estate. When the feed is empty and your connectors have been checking in, the dashboard says so and tells you when they last checked.

That message is careful. If a connector is failing, or has stopped polling, an empty feed is not an all-clear, and the dashboard says exactly that instead of reassuring you. The same goes if the connector list itself could not be loaded.

Each row shows:

A coloured dot for severity: critical, warning, OK (a recovery) or info.
A title and, underneath, the detail.
The connector that raised it, as a small tag, and how long ago.
Buttons for Why?, Acknowledge and Snooze. Critical events also get Draft Comms.

Rows can carry small badges beside the title:

Resolved -- a later recovery event has ended this problem. The title is struck through. Click the badge to jump to the recovery.
Superseded -- a newer event for the same condition is the current one, for example a warning that has since become critical. It is tagged but not struck through, because the problem has not gone away.
Maintenance -- the event fell inside a maintenance window, so it was recorded as expected rather than as a fault.
Acknowledged or Snoozed -- see below.
2nd time in 30 days (or 3rd, 4th and so on) -- see the next section.

Use the buttons above the feed to show only Critical, Warning, OK or Info events, or to Hide resolved ones. Your choice for hiding resolved events is remembered in your browser. The feed shows fifty rows at a time; click Load more for the next fifty.

Problems that keep coming back #

A problem that is resolved is not necessarily dealt with. When the same condition has been raised more than once in the last 30 days, its row says so: 2nd time in 30 days, 3rd time in 30 days, and so on. Hover over the badge for the full sentence. The count includes struck-through rows. A first occurrence shows no badge.

This is the quickest way to tell a one-off from something that needs a proper fix. The Recurring Events panel on the Intelligence page (see AI Features) looks at the same question across the whole estate.

How long a problem has been open #

The dashboard does not show the age of a problem. The SLT Dashboard does: beside each item on an amber or red card it says how long the problem has been going on.

Acknowledge and snooze #

Both are for tidying the feed while you deal with something. Neither fixes anything, and neither stops Groundskeeper noticing a problem.

Acknowledge dims the row and tags it Acknowledged. There is no timer. Click Unacknowledge to bring it back.
Snooze dims the row and tags it Snoozed for 15 minutes, 1 hour, 4 hours or until tomorrow (8am). When the time is up the row returns to normal by itself. Clear / reactivate ends a snooze early.

The Alerts page has the same two buttons and takes an event off its list when you use them. The SLT view's lights are not affected.

Admin and technician accounts can both acknowledge and snooze. Nothing is deleted.

Working on one connector at a time #

Click a connector in the status panel on the right, or its small tag on any feed row, and the feed shows only that connector. Click it again to clear the filter. The filter also pulls in that connector's older events, so a quiet connector is not hidden just because busier ones have filled the feed.

While one connector is selected, two extra buttons appear:

Acknowledge all -- acknowledges every outstanding event from that connector. It can be undone one row at a time.
Clear... -- permanently deletes that connector's events. Admins only, and it asks you to confirm first. The latest alert for each problem that is still open is kept, so a failure that is still happening stays on the dashboard until it is fixed. If you only want less noise, use Acknowledge all instead.

Why? and Draft Comms #

These two buttons need AI to be switched on. With AI off, they say so.

Why? asks the AI to explain that event in plain English. The answer opens under the row. Click Close to hide it. It stays open when the feed refreshes.
Draft Comms (critical events only) drafts a short note for senior leaders, with a Copy button so you can paste it into an email and adjust it there.

Both count towards the daily AI question limit described on the AI Features page.

The AI summary and Ask Groundskeeper #

When AI is on and a summary has been made, a short banner at the top of the page sums up the state of the estate. It is made in the background, so the page never waits for the model. If AI is off, there is no banner.

Under it, the Ask Groundskeeper box takes a free-form question ("Is everything OK today?"). The full conversation link opens the Intelligence page, where you can ask follow-up questions. With AI off, the box says so.

Search and export #

The search box above the feed finds events whose title, detail, connector or device mention what you type. It searches the whole feed history, not just the rows on screen, so "every event mentioning FS01" works even if FS01 has been quiet this week.

Export CSV downloads the last 30 days of events as a spreadsheet. If you have typed something in the search box, only the matching events are exported. Each row has the time (UTC), severity, connector, title, detail, device, and when it was acknowledged or until when it is snoozed. It opens correctly in Excel. Admin and technician accounts can export, so it suits sending to a governor or an insurer.

Connector status #

The panel on the right shows how each connector is doing.

At the top, a one-line summary such as "N of M connectors healthy". A connector that has stopped polling does not count as healthy.
Each row shows the connector's name, when it last polled, and its state: OK (green), Error (red), or Not polling (amber -- it has not polled successfully for longer than expected, though nothing has failed).
The System entry gathers Groundskeeper's own housekeeping: the latest backup, a restore waiting to apply, an update that is available, and any maintenance window that is active now. Click it to see those events.
Once agents are set up, an RSJ Agent entry shows how many enrolled devices are online. Click its status to open Agent Settings.

If a connector shows an error, its row in the feed says what went wrong. Open Connector Settings to check its credentials and settings. The Connectors page explains each one.

Banners you may see #

Some dashboard data isn't loading -- Groundskeeper could not reach part of its own service. Use the link to open the health check.
Running as Local System, but a service account is configured -- an upgrade reset the service's log-on account. Open General Settings -> About and click Apply now. See Service Account.
Getting started -- a short checklist on a new install. It ticks itself off, and you can dismiss it.