Docs

Status, Processes and Resources

Watch the server's services, load, disks, processes and background tasks, see what each hosting account uses, and stop programs that should not run.

12 min read Updated From the panel's guide

This page covers two sections of the admin menu. Status and Monitoring shows the state of Apache, the services, the hardware and the panel's background tasks. Processes and Resources shows what runs on the server and what each hosting account uses, and lets you end programs.

Which services the panel starts with the server and watches is set under Services and Monitoring, described in Software and Service Settings. Email alerts for failed services and full disks go to the address under Alert Contacts.

Apache Status#

In the admin area, open Status and Monitoring → Apache Status (/admin/tool/apache-status). The page shows Apache's own status report, read on the server itself.

  • What Apache is doing counts Apache's workers by state, for example how many are waiting for a connection, reading a request, sending a reply or keeping a connection open.
  • Figures lists the values Apache reports about itself, such as its uptime, the requests and data served, and the number of busy and idle workers.

If Apache does not report its status, it may not be running; check it under Service Status.

Use it to see whether Apache is running out of workers when websites become slow. If most workers are busy most of the time, raise MaxRequestWorkers under Main Apache Settings or look for the websites causing the load in the Process Manager.

Good to know:

  • The status report is not reachable from outside the server; only the panel reads it.
  • Switching on ExtendedStatus under Main Apache Settings adds more detail about each request.

Daily Process Report#

In the admin area, open Status and Monitoring → Daily Process Report (/admin/mtool/daily-process-log). The panel measures what every user on the server uses every five minutes and keeps the results for 31 days. Use it to find the accounts that load the server.

The page shows one day at a time. Use the links at the top to go to the previous or next day.

  • Averages (since midnight, or of the chosen day) lists each user with their average % CPU, highest % CPU, % memory, memory in MB and number of processes. Click a column heading to sort. Hosting accounts link to their account page.
  • For today, two more tables follow: Usage by user, right now and Busiest processes, with the user, % CPU and the full command of each process.

The page says how many measurements the day has and how many processors the server has. 100% CPU means one processor fully used, so on a server with four processors a user can reach 400%.

Good to know:

  • CPU figures of the busiest processes are each program's average over its running time.
  • To end a program, use the Process Manager.

Server Information#

In the admin area, open Status and Monitoring → Server Information (/admin/mtool/server-info). The page describes the server's hardware and system:

  • Processor: the number of processors (threads), vendor, model, speed, cache, cores per processor and BogoMIPS.
  • Memory: total, free and available memory, buffers, cache, swap total and free, and similar figures.
  • System: the operating system, hostname, kernel, how long the server has been running and the load averages.
  • Disk drives: each drive with its model (when the drive reports one), whether it is a hard disk or an SSD or virtual disk, and its size.
  • Disk space: each file system with its size, used and available space, percent used and mount point. File systems that are 90% full or more are marked.

The page only reads; nothing can be changed here. More detail about disks is under Disk Usage; running programs are in the Process Manager.

Service Status#

In the admin area, open Status and Monitoring → Service Status (/admin/mtool/service-status). This is a quick health check of the server on one page.

  • Services lists the services that are being watched, with their version and state: "up", "starting" or "down" (with the reason the system gives). Which services are watched is set under Services and Monitoring; a service that is not watched does not appear here.
  • Load and memory shows the load of the last minute next to the number of processors, the memory in use and the swap in use. The load is marked "high" when it is more than twice the number of processors, and memory when 90% or more is in use.
  • Disks lists each file system with its usage. It is marked "filling" from 80% and "full" from 90%.

To restart a service that is down, open its page under Restart a Service in the menu.

Good to know:

  • The panel checks watched services every five minutes on its own and emails you when one has failed, whether or not this page is open.

Background Task Queue#

In the admin area, open Status and Monitoring → Background Task Queue (/admin/mtool/task-queue). Long tasks run in the background, one at a time: backups, transfers, restores, system updates, package and PHP installs, certificate runs, database checks and similar work.

The page shows the task running now with its progress and log, or, when nothing runs, the last task that finished and how it ended. While a task runs, the page refreshes itself.

If no long task has run since the panel was started, the page says so.

Good to know:

  • Because tasks run one at a time, starting a second long task while one runs is refused; wait until the first has finished.
  • The page that started a task shows the same progress at its top, so you do not have to come here to follow it.

Forbidden Background Programs#

In the admin area, open Processes and Resources → Forbidden Background Programs (/admin/mtool/process-killer). The panel can end certain programs whenever a hosting account runs them, and email you each time it finds one. IRC bouncers, bots and coin miners are what misused hosting accounts typically run; they tend to bring abuse reports and attacks.

The programs you can choose are BitchX, bnc, eggdrop, generic-sniffers (sniffit, dsniff, ettercap, linsniffer), guardservices, ircd, psyBNC, ptlink, services and miners (common coin-mining programs such as xmrig, minerd, cpuminer, kdevtmpfsi and kinsing).

To set it up:

  1. Tick the programs to stop. Ticking all of them is recommended.
  2. Optionally, under Users allowed to run them, enter account usernames (one per line) that may run these programs anyway.
  3. Click Save.

Good to know:

  • The check runs every five minutes.
  • Programs of root and of the server's own services are never touched; only hosting accounts are checked.

Process Manager#

In the admin area, open Processes and Resources → Process Manager (/admin/processes). The table lists the processes running on the server, busiest first, measured when the page loaded: process ID, user, priority (nice, from -20 most favoured to 19), processor %, memory % and MB, how long it has run, and the command.

  • Processes of: choose a user and click Show to see only their processes. Clicking a user name does the same.
  • Details shows what a process is: its command, program, working folder, state and what it waits for, threads, the process that started it, memory now and highest, disk reads and writes, control group, and its open files and connections.
  • End asks the program to stop, letting it finish its work.
  • Kill stops it at once.
  • Kill every process of the user: choose a user and click Kill them all to stop all their processes at once.

Good to know:

  • Killing a process can lose data it was writing. Try End first.
  • Root's processes are the system itself: they cannot be killed all at once, only one at a time. Process 1 and the panel's own processes cannot be ended from the panel.
  • To end a known bad program automatically every time it appears, use Forbidden Background Programs.

Resource Usage#

In the admin area, open Processes and Resources → Resource Usage (/admin/mtool/resource-usage). The page shows what everything running as each hosting account (PHP, scheduled tasks, shell sessions) uses right now, next to the limits of its package. The limits are processor, memory, disk throughput and number of processes, set under Packages.

The heading says whether the limits are active. For each account the table shows the package, processor %, memory in MB, disk MB per second, processes, and how often a limit was hit. Click History to see charts of the last 24 hours (one point per minute) for processor, memory, disk and processes, with the account's limit as a dashed line.

If the server was started with the older kind of control groups, the limits cannot work. The page then offers Switch to unified control groups: click it, then restart the server under Restart the Server Safely to switch the limits on.

Good to know:

  • A program that reaches the memory limit is ended by the system; one that reaches the processor or disk limit only runs slower.
  • If the page says the service that holds the limits is not running, check systemctl status intopanel-accounts on the server.

Disk Usage#

In the admin area, open Processes and Resources → Disk Usage (/admin/mtool/disk-usage). The page shows the server's disks from two sides.

  • Disk space per file system: device, file system type, size, used and available space, percent used, the percentage of files (inodes) used, and the mount point. Click a column heading to sort. File systems 90% full or more are marked. A file system can also be full because it has run out of inodes (too many small files), even with free space left.
  • Disk activity: per device, transfers per second, blocks read and written per second, and the total blocks read and written. These are averages since the server was started; a block is 512 bytes.

What each hosting account uses on disk is shown under All Accounts. Accounts at their limit are under Accounts Over Their Disk Limit.

Running Processes#

In the admin area, open Processes and Resources → Running Processes (/admin/mtool/running-processes). This is a plain list of every process on the server, with its process ID, program name, user, the executable file it runs from, its working folder and the full command.

Use it to find out where a program lives and what started it, for example a suspicious program running from a website's folder or from /tmp.

The page only reads. To end processes or see their processor and memory use, open the Process Manager.

Server Monitoring#

In the admin area, open Add-ons → Server Monitoring (/admin/monitor). intoPanel measures the whole server once a minute and keeps eight days of measurements in /var/lib/intopanel/monitor (one file per day). Nothing is sent to an outside service. The first measurements appear a minute or two after the panel is installed or updated.

Choose the period at the top: one of the ready periods (last hour up to last 7 days), or From – to with your own dates and times and Show. For a ready period the page loads the new measurement by itself after "Next update at"; the ↻ link loads it at once. Move the pointer over a chart to read the values at that moment.

  • Overview: CPU, memory (used, buffered, cached, free), disk reads and writes, load average (1, 5 and 15 minutes), network bits in and out, and the programs that used the processor most.
  • Network: response times of three outside addresses (a connection to their DNS port; a gap means no answer), bits and packets per second for the whole server, and the traffic and packets of each network interface.
  • Memory: memory use as shares and in bytes, available, active and inactive memory, and swap.
  • CPU: CPU usage, the use split into user programs, system, waiting for disk and steal (time a virtual server's host gives to others), and the load average with the number of cores.
  • Disk: the space of all local file systems over time, reads and writes, and a table of every disk mount with its file system, percentage used, bytes used and bytes free.
  • Processes: the top programs over the period, and the top 10 processes right now (measured over one second when the page opens) with parent process ID, PID, user, program, CPU, memory and bytes read and written. Sort the table by clicking a heading.
  • Alerts: the alert history, filtered by status (open, closed) and severity (informational, warning, critical), and the thresholds.

An alert opens when the server passes a threshold: the processor busy for five minutes, the five-minute load per core, too little memory available, a file system too full, swap too full while memory is short, or no outside address answering for three minutes. It closes by itself when the server is back under the threshold; Close closes it at once (it opens again if the threshold is still passed). A start of the server is noted as informational. Enter 0 as a threshold to leave that check out.

A full swap alone raises no alert: Linux moves memory that has long been idle to swap and leaves it there while RAM is to spare. The swap alert opens only when, at the same time, the available memory is below one and a half times the memory warning (at least 15%), and both stay so for 15 minutes: backups, restores and the nightly malware scan fill swap for a while, which is no trouble. When Add a notification is ticked, every new alert also appears under the bell and is mailed to the alert contacts.

Home Screen Dashboard card switches the Server Monitoring card on the home dashboard on or off. It shows the processor, memory in use, the fullest disk, the load and the open alerts.

Optional Software: PostgreSQL and Calendars and Contacts#

PostgreSQL (with phpPgAdmin) and Calendars and Contacts (Radicale) are installed with intoPanel. In the admin area, open Software → Optional Software (/admin/software) to see them, and to remove one you do not want: tick the box and press Remove. Its tools disappear from the client area, and updates of the panel leave it out until you press Install here again. Removing keeps the data (PostgreSQL's databases in /var/lib/pgsql, the calendars and address books in /var/lib/intopanel-dav), so installing again brings it back.

PostgreSQL is the version the operating system provides (16 on AlmaLinux), with the PostgreSQL extension for every installed PHP version, and phpPgAdmin at /phppgadmin on the server's hostname and every hosted domain; the client area opens it on the hostname over https, as it does phpMyAdmin. It listens on the server only and checks passwords with SCRAM-SHA-256; the database administrator (postgres) cannot sign in to phpPgAdmin, and phpPgAdmin shows each user only the databases it may use. Accounts use Databases → PostgreSQL in the client area, unless their package leaves out "PostgreSQL Databases". Account backups include a dump of each PostgreSQL database, restored by hand with pg_restore.

Calendars and Contacts (CalDAV and CardDAV) runs on port 2080 (https, with the mail server's certificate), as the service intopanel-dav. The installer opens port 2080 in the firewall once. People sign in with a mailbox's address and password, checked by Dovecot; each mailbox sees only its own calendars and address books, and the mailboxes of accounts whose package leaves out "Calendars and Contacts" cannot sign in.

Was this article helpful?

Something missing? Ask in the community or open a ticket.