Overview

The first page you see after logging in. It answers four questions: what is this machine, is it healthy, how many resources are left, and who has done what recently. The page has two sub-pages: Summary and Monitor. Only accounts with "host read-only" permission get Monitor; accounts without it see only the summary of virtual machines.

1. Summary

1.1 Top: host name, status and three gauges

On the left is the host itself:

Host name: the large text is the host name, the small text the domain name. Accounts with host write permission have a pencil button next to it (Change host name); the dialog asks for Host name and Domain (the domain may be empty). Click Save and it takes effect immediately. If the management network is not managed by keel, or a network change is being tested, you are told it cannot be changed. When you use your own HTTPS certificate, the dialog tells you whether the new name is among the names in the certificate.
The machine's manufacturer and model.
Version: the keelOS version currently running; click it to go to Upgrade and restart.
Status: green "OK" when there is nothing wrong; when something needs attention it shows N to look at, with the details in §1.2.
Uptime: counted from the last boot or reload.

On the right are three gauges, each with a percentage, a usage bar and the amounts: CPU (usage at the latest sample and the number of threads), Memory, and Storage (used, free and total of the default storage pool). The bar turns red above 90%. On hosts where keelOS manages virtual machine memory itself, the Memory gauge shows the part reserved for virtual machines; the memory of the whole host is in the hardware table below.

1.2 Things that need attention

A row of notice bars may appear below the top; each is one thing to deal with, and clicking it jumps to the relevant page. Common ones:

No storage pool yet (create one in Storage); a storage pool that is not ONLINE.
A network port has no link (only reported when that port is in use by the host address, a switch or a network; an idle port does not count).
A restart is needed (upgrade or pending change); see Upgrade and restart.
A virtual machine has failed or has a damaged configuration.
A memory module reported errors: uncorrectable errors mean the module needs replacing, and correctable errors within the last 24 hours are reported.
A passed-through device's interrupt cannot be delivered; see Device passthrough and hardware.
A manager is asking to manage this host; click in to approve or refuse.
Configuration problems: outdated or problematic settings found by keel, each with a Fix button. At most 6 are listed; the rest are summarized in one entry that links to Configuration → Problems.
The virtualization component used by virtual machines could not start (for example AMD-V is turned off in the firmware); when the host is not running on keelOS's virtual machine monitor there is another notice, an informational one (not a warning).

When the host enters Safe mode (when the keelRM management port detects a problem), a separate notice card gives the reason and a Leave safe mode link; see The keelRM management port.

1.3 The hardware and system tables

Hardware (click details at the top right to go to Configuration → Hardware): Manufacturer, Model, CPU (model, number of sockets, cores and threads), Memory (the total, and the part reserved for virtual machines), Network ports (a small tag per port; several ports of the same card are joined together; green means link is up; click to open the port page), Storage pools (name, state, size).

System: Host name (with a clickable link beside it if there is an mDNS name), Address, Gateway, DNS, Host time (with time zone), System (the underlying FreeBSD version), Serial number, BIOS (version and date). Items that do not exist are not shown.

1.4 Virtual machine summary

The Virtual machines card gives the number in each state (running, stopped, and where applicable suspended and failed), followed by small tags for the running virtual machines: name, number of vCPUs and memory; click one to open its detail page. With more than 16 machines it shows +N, linking to the list.

1.5 Performance (last hour)

Performance, last hour has several line charts, each with now · average · peak next to its title:

CPU use;
memory: Memory in use (on hosts with a virtual machine memory pool, Guest memory in use);
network (physical ports): receive, send;
Disks: Read, Write.

Move the mouse over a chart to see the value at a moment. Right after boot it says data is being collected, and a few seconds later there is data. Each virtual machine's own CPU, memory, network and disk curves are on its detail page; see Virtual machines.

1.6 Recent activity

Accounts with host write permission also see Recent activity: the 8 most recent operations (time, action, object, source address, result), with failed results marked orange. Click Audit log to see all of them; see Accounts, tokens and permissions.

2. Monitor

The Monitor sub-page holds the details that are not part of the summary. From top to bottom:

CPU: one cell per core, showing whether it currently belongs to the host or to which virtual machine, big or small core, the distribution across NUMA nodes, and how much memory reserved for virtual machines is left. On multi-socket hosts it is how you judge whether virtual machines are placed well.
Memory bandwidth (last hour): the reads and writes of each CPU's memory controller (virtual machines and host together). Only available on CPUs whose counters can be read.
Temperature, Fans: curves over time of the temperature sensors the host can read (CPU, ACPI thermal zones, PCH, disk SMART) and the fan speeds. If no sensors can be read, it says so.
Memory: zero pages: keelOS periodically looks for whole blocks of zeros in virtual machine memory and reclaims them (reclaimed memory still counts as that virtual machine's, so it does not get less). It lists the usage per virtual machine and shows how much has been lent to the host as cache.
ZFS cache (ARC / L2ARC): the size and hit rate of the ZFS read cache, the usage of the L2ARC cache disk and the current tuning parameters. Counters count from the host's boot.

Note: the Monitor sub-page is read-only; you cannot change any setting from here.

← Previous: Creating your first virtual machine Next: Virtual machines →