Skip to main content

Reports

The Reports module turns the raw monitoring into management-grade answers: are we keeping our promises, what is drifting, what are we overpaying for, and who gets the summary in their inbox.

SLA and error budgets​

Define SLO targets on what matters: a host, a service, or a site, with a target percentage over a rolling 30-day window. The target translates into an error budget (the minutes you are allowed to be down), and each objective shows its uptime, the budget consumed, and the minutes left, colored honestly from healthy to burned.

When an objective has eaten 80% of its budget, a budget burn warning incident fires, so you hear about a promise in danger before it is broken, not after. The incident says what is true: "SLO budget running low" while the target is still met, "SLO target missed" once the budget is gone, and it names the host. Available on Business and above.

SLA fairness​

Three details keep the numbers honest. An objective belongs to a concrete subject (this host, this service, this site), so the report never averages away the thing you promised. The denominator is the minutes the host actually owed: from its first heartbeat, up to the last completed minute, minus the minutes the platform itself was unreachable or restarting. And an operator can declare excluded periods on a host, windows when it was knowingly not being observed (a migration, a lab power-down): those minutes leave the availability math entirely instead of counting as downtime.

Weekly digest​

A periodic email for humans: the period's KPIs compared against the previous one, what is still broken (open incidents, not just new ones), the noisiest offender, fleet growth, and what is coming up (certificates and domains about to expire). The window follows the schedule, so a monthly send reports the month, not its last week. Recipients and the schedule are configurable, and send test delivers one right now so you can see it before committing anyone else's inbox. Available on Pro and above.

Report builder​

Build a custom report instead of taking a canned one: pick the scope (the whole fleet or one host), the sections to include, the period, and the language the report should speak. The result renders as a clean print sheet with a live preview, exports to PDF, and can be handed to a client as is. A single-host report opens with a verdict block: what is fine, what needs watching, what is overdue, including operating system and service releases checked against their published end-of-life dates, so "this box runs a release that no longer receives fixes" arrives as a dated fact.

The sheet speaks the client's language, not ours: it reports incidents, availability and findings about their systems, never our collection mechanics.

Capacity and waste​

The other direction of monitoring: what you are paying for and not using.

  • Idle hosts: boxes that do nothing at all over the window: low CPU and memory, no traffic to speak of, no requests, and no monitored services beyond the operating system's own. A quiet box that still serves a site or holds a database is not idle, so the list is a consolidation shortlist you can act on. Each entry carries an honest "this host costs N per month on your plan" where that is actually true (inside your plan's quota it says so instead of inventing savings; agency plans price the box at the wholesale rate it really costs).
  • Rightsizing: services that want a different size, like PHP-FPM pools pinned at their ceiling, or hosts with large unused memory headroom.
  • Energy and carbon: an honest estimate of the fleet's power draw and CO2 footprint, derived from core counts and real CPU load with stated assumptions (power per core, datacenter overhead, grid intensity), all of which you can adjust to your region. The core count printed is the one a host has now, not an average over the window, so a box that lost or gained cores reads correctly the same day; the watts follow the window's average load. It is labeled as the estimate it is, and it makes the idle-host list read in kilograms as well as euros.
  • The dashboard carries a compact cost-and-waste card with the same numbers.

Available on Pro and above.

Scheduled exports​

Any report can leave the building on a schedule: pick the report, the format (PDF or CSV), weekly or monthly, and the recipients, or hit Run now. Schedules keep the timezone of whoever saved them, so "Saturday 09:00" means your Saturday morning, not a server's. PDFs are print-clean, and under an agency's white label they carry the agency's brand. See For agencies.