Open app

Dashboard & Actions

The dashboard answers one question: what needs my attention today? It ranks your fleet by what’s wrong, not alphabetically by hostname.

The header

At a glance you see the number of servers observed, how many are critical, aggregate Health, coverage, pending security updates and servers needing a restart.

Actions

An Action is something Demerzel thinks you should look at — derived from the evidence, ranked by severity, with what to do about it. The kinds you will meet:

AreaTypical Actions
Updates & lifecycleSecurity updates waiting; a restart needed to finish updates; a new kernel installed but not running (with the CVEs the reboot would fix, when known); critical or high vulnerabilities with a fix available; an operating system or runtime past its end of life
CapacityA disk nearly full, or filling fast enough to run out soon; sustained memory pressure; a server that looks oversized (cut cost) or undersized (add CPU)
AvailabilityFailed services; an agent that went quiet — with whether the host itself still answers (see Notifications)
ExposureA database or admin service reachable from the internet; SSH exposed with no fail2ban protection; domains served without TLS; a certificate about to expire
Web hardeningA site serving a secret file (.env, .git…); phpMyAdmin with no basic auth in front; more PHP workers than the database accepts connections; no slow query log; sites behind Cloudflare that log Cloudflare instead of the visitor; fail2ban guarding SSH only while the web server takes scanner traffic
Security incidentsAn attack that hurt the service — errors served to real visitors — linked to its incident report
Active DefenseThe warden frozen by its cap, not running, with a broken policy, or unable to receive the agent’s decisions; a rule demoted after it erred; an address spared because it carries too large a share of your traffic

Each Action gives you three ways to respond:

ControlWhat it does
AcknowledgeRecords that a human has seen it — it stays visible but marked.
SnoozeHides it for a while (1 hour, 1 day, 1 week), then it comes back if still unresolved.
DismissHides this occurrence until the situation materially worsens.
Directional dismiss. Dismissing disk usage at 82% won’t re-alert at 83%, but it will reopen at 94%. You silence noise without going blind to a real deterioration.

Every decision is journaled — see Activity log & audit. Read-only members can view Actions but not act on them.

Recently resolved & recent changes

Below the Actions, the dashboard shows what resolved itself in the last week and a feed of recent attribute changes across your fleet, so you can see movement without digging.