Monitoring

See alerts and health across all your hosts from one dashboard.

The Monitoring area gives you a fleet-wide view of alerts across all your hosts, so you can spot problems without opening each host individually. For teams running business-critical sites, this helps you notice and react to problems quickly - for example during a busy campaign - before they affect visitors.

The alerts dashboard

Alerts are grouped by priority - from P1 (most urgent) down to P4 - and sorted so the most critical issues are always at the top. They are split into four categories:

Alerts - the main server and application checks in a non-OK state.
Mail Problems - mail-specific issues, such as a growing mail queue or Sender Policy Framework (SPF) or Domain Name System (DNS) warnings.
Security Issues - hosts with security or system packages that still need updating (for example an apt check reporting critical updates to apply).
Other - lower-priority signals worth reviewing but not urgent.

Each alert names the affected host and service (for example a disk, swap, certificate, or cache-hit-ratio check) with a short status message and how long it has been active. Selecting an alert row opens the alert detail page for that check. The host name in the row is a separate link and opens the host instead. The dashboard refreshes as new alerts arrive.

The Monitoring dashboard with the Critical, Warning and Unknown filters and the Alerts, Mail Problems and Other panels

The alert detail page

Selecting an alert opens a page for that one service check, at /monitoring/<host>!<check>. The heading shows the host name followed by the check name, and the host name links to the host. You also reach this page from the CPU, RAM and Disk graphs on a host's Health tab, which open it in a new browser tab.

The page reads the current state from monitoring every time you open it. It has three panels and one action:

Status - the current state and the check output
Graph - the measured values over time
Incident reports - reports written about this check
Create a ticket - hand the alert over to support

Select a panel header to collapse or expand it.

Note

If the check no longer exists in monitoring, the page reports that it cannot be found. Go back to the dashboard and pick a current alert.

Status

A badge next to the panel title shows the current state: OK, WARNING, CRITICAL or UNKNOWN. The panel lists:

Field	What it tells you
Since	How long the check has been in this state, with the date and time it changed.
Groups	The monitoring groups this check belongs to.
Plugin Output	The message the check returned. It usually names the exact problem, for example a filesystem that is nearly full or a certificate that is about to expire.

Each field is shown only when there is a value for it.

Graph

The Graph panel plots what the check measured. Use the 1H / 8H / 1D / 7D buttons to change the time range, from one hour to seven days. The panel is not shown for checks that record no measurements, such as a check that only reports a state.

Tip

Compare the 1-hour and 7-day views before you act. A single short spike is usually harmless; a line that climbs day after day needs attention.

Incident reports

When incident reports exist for this host and check, they are listed here, with the number of reports next to the panel title. Each entry shows the report title, the alarm it relates to, and how long ago it was written.

Open report shows the full report in a window on top of the page.
PDF opens the same report as a Portable Document Format (PDF) file in a new browser tab, where you can read or save it. The button appears only when a PDF version exists.

The report window has the same PDF button in its header. Select Close or the cross in the header to return to the alert.

Create a ticket

Create ticket, at the top right of the page, opens the Hosted Power ticket portal in a new browser tab. The subject is filled in with the host name and the check, so you do not have to retype it. Use it when an alert needs follow-up from the support team - see Support.

Per-host health

For the live health of a single host - top issues, CPU/memory/disk usage and individual service checks - open that host's Health tab.

Understanding the metrics

The per-host view (and the Health tab) report four core resource metrics:

Metric	What it means
CPU usage	How busy the processor is. Sustained high values mean the host is CPU-bound.
RAM usage	Used memory. Consistently high RAM with rising swap points to memory pressure.
Memory swap	Memory spilled to disk. Steady swap use hurts performance - a sign to investigate.
Disk usage	How full each filesystem is. A full disk causes failures; act before it fills.

Each service the host runs (web server, databases, Redis, search, mail, certificates, and more) is also checked individually and reports OK / Warning / Critical.

Alert priorities

Alerts are grouped by priority so you can triage quickly:

Priority	Meaning
P1	Critical - something is down or broken and impacting the server. Needs immediate action. Max resolution time: immediate.
P2	Very urgent - down, broken or degraded. Max resolution time: 1 hour.
P3	Urgent - a problem that could escalate if left unhandled. Max resolution time: 4 hours.
P4	Not urgent - worth reviewing for long-term stability. Max resolution time: 48 hours.

Important

Not every alert is equally urgent, and not all of them need action. Triage them in order:

P1 and P2 first - these affect, or are about to affect, your live sites. Act on them right away.
P3 next - look at these when you have time; usually a trend to watch rather than an outage.
P4 is optional - informational only, fine to review during normal maintenance.

Working top-down (P1 first) keeps your attention on what actually impacts visitors.

Alerts are also grouped by type, such as Mail Problems and Other. Hosted Power's team monitors these alerts; follow up through Support for anything affecting your sites.

How monitoring works

TurboStack monitors your infrastructure and applications, and can integrate application performance monitoring. For the underlying tools and what each provides, see Monitoring (concepts).