Skip to documentation
Interneting.orgDocs

Documentation

Performance

Inspect resource charts, memory cache and current or historical Site use.

Server → Performance shows host, service and Site resource use. Use it to investigate pressure and compare recent periods without treating historical samples as live readings.

Choose a time range

Select the period you want to inspect. Historical CPU, RAM and swap averages/peaks follow that range. Since Update covers the samples retained since the current Control release was installed; if the retained history is shorter, Control shows only the available samples.

Current and sampled values

Current host RAM, swap and disk figures are read from the server when the page is loaded or refreshed. CPU history and some Site/service figures come from periodic samples and show their observation age where relevant. A value labelled as a sample is not a new live reading.

Host CPU and memory

Use CPU, RAM, swap, disk and resource-pressure indicators to see when the server is busy. Available RAM is usually the most useful capacity guide because Linux reuses spare memory for filesystem caching and can reclaim much of that cache when applications need it.

The memory breakdown also separates Sites, shared services, Control and other host processes where reliable attribution is available. Do not add unrelated cache limits and process-memory figures together as though they represented the same memory.

Infrastructure

Infrastructure shows current process/service observations such as ingress and database memory. Pressing Refresh requests a new observation; rapid range-only changes may reuse a recent process sample and display its age.

Site performance

Open a Site and find Performance for its CPU, memory, disk I/O and traffic history. A Site's historical averages/peaks follow its own period selector. Current or fixed-window values are labelled separately.

Resource Governor

The Governor shows whether Control can admit more background work without unnecessarily competing with running Sites. Waiting work can be caused by host capacity or by another operation holding a shared resource.

Missing readings

After a restart or new installation, some metrics need additional samples before a history or rate can be calculated. Missing data is shown as unavailable/collecting rather than as zero.