PanelicaDocs
panelica.com
Docs / Monitoring / Resource Usage

Resource Usage

Panelica 1.0.375 Verified 2026-07-24 Monitoring

Resource Usage shows what every account is actually consuming, live, against its plan limits: CPU, memory, disk, inodes, disk I/O and process count, per user. The numbers come straight from the kernel's own accounting (cgroups v2 and filesystem quotas), so they reflect real enforcement, not estimates.

Panel locationMonitoring → Resource Usagehttps://YOUR-SERVER-IP:8443/resource-usage
 Resource Usage
Total Users
12
Quota Exceeded
1
High Usage
2
Normal
9
UserCPURAMDiskPIDStatus
john12% / 100%340 / 512 MB4.2 / 5 GB18 / 50Normal

What it measures

The page is per user, not per domain, because Panelica isolates resources at the user level: all of an account's domains share its cgroup slice. Four cards count total users, quota-exceeded, high-usage (any resource at 75%+) and normal accounts. The table then shows, for each user, their plan, and live: CPU (percent of their limit), RAM (used/limit), disk (used/limit), inodes (used/quota, when kernel quotas are active), IO read/write rates, process count (used/limit), the plan's quota mode, and a status badge. Bars turn amber at 75% and red at 90% of the limit, and the busiest accounts sort to the top automatically.

Behind the scenes: the panel reads each account's cgroup files directly (cpu.stat, memory.current, io.stat, pids.current) and its filesystem quota counters via a kernel syscall (no slow du where quotas are on), computing rates from deltas between reads. Results are cached for a few seconds and served stale-while-fresh, so the page stays responsive even with hundreds of accounts. Refresh interval is adjustable (1 to 20 seconds) with a pause toggle.

What "quota exceeded" means here

The status badge is a live indicator, and it is worth understanding exactly what enforces the limit, because it is not this page.

Enforcement is in the kernel, always on, independent of this display. The "Quota Exceeded" badge is a reading, not the mechanism. The real limits are enforced by the kernel continuously: CPU over the plan limit is throttled by cgroups (not stopped), RAM over the limit triggers a kernel OOM-kill of the offending process, the process cap blocks new forks, and disk/inode limits stop new files from being created (while the site keeps serving and deletes always work). These happen whether or not you are looking at this page, and whether the plan is strict or not, because the limits are programmed into the kernel when the account is provisioned. Alerts fire to administrators when thresholds are crossed.
Reading the numbers honestly: Linux counts CPU as a share, so a single-core plan showing 100% is fully using its allocation, not the server. RAM "used" excludes reclaimable cache. The IO columns are byte-per-second rates, not totals. And because enforcement is per-user, a customer with several busy sites sees them all summed into one line here, which is exactly how their cgroup limit applies.

Access

Behind the resource_usage_dashboard license feature (Pro and up). Visibility follows the roles and permissions chain: a user sees only themselves, a reseller their customers, an admin their chain, ROOT everyone. The limits shown are the account's plan limits; how those limits are set lives in Hosting Plans.

Panelica Documentation · Written and verified against the live panel. · Last verified 2026-07-24 on Panelica 1.0.375