Skip to content

Platform Status

Platform Status gathers the health of your platform onto one page: what the billing queue is doing, how your users are signing in, how much of your activity comes through the API and AI assistants, and how the server itself is coping.

It is the page to open when something feels slow, when a billing run seems not to have started, or when you want a periodic check that your users’ security is in order.

Platform Status is a tab on the main menu, to the right of Advanced. It needs elevated access with two-factor verification. If the tab is not there, ask your platform administrator to check your access level.

The page loads its figures when you first open the tab, and holds them. Clicking away and back does not fetch new ones, so the numbers are a snapshot from when you opened it, and Platform Time tells you when that was. Reload the page for current figures. This matters most for the queue counts, which move constantly.

Some panels need access beyond the tab itself, so you may not see all of them:

PanelWho sees it
Platform & Queueanyone who can open the tab
User Access & Security, API & MCP Activityusers with data-protection report access for account status
Server Information, OPcache and APCuplatforms running on their own dedicated server

The billing task queue and the platform’s own mode and clock.

RowWhat it tells you
Platform ModeWhether this is your live platform or a test one
Platform TimeThe server clock, and effectively when these figures were taken
Ready TasksTasks waiting to run now
Delayed TasksTasks deliberately scheduled for later
Running TasksTasks in progress
Completed with Errors (24h)Tasks that finished but reported problems
Failed Tasks (24h)Tasks that did not complete
Oldest Ready TaskHow long the longest-waiting task has been queued

Oldest Ready Task is the one to watch. Ready tasks are normal - work arrives and is picked up - but a ready task that has been waiting a long time means the queue is not being worked through, and anything you have just requested is behind it. A steady rise in Ready Tasks alongside it points the same way.

Failures and errors in the last 24 hours are a prompt to look at Billing Tasks itself, where you can see which task failed and why.

Counts of your users and how they have been signing in. Everything here is a count or a timestamp; no individual user is named. System users are left out throughout.

RowWhat it tells you
Active UsersActive accounts on the platform
Active Login UsersOf those, the ones that can sign in
Active Elevated UsersOf those, the ones holding elevated access
Users Logged In (24h) and (30d)How many distinct users actually signed in
Latest Successful LoginWhen the most recent sign-in happened
Failed Login Events (24h)Unrecognised users, wrong passwords, expired passwords, inactive accounts and failed two-factor attempts
Platform Views (24h) and Platform Edits (24h)Activity through the interface, excluding API and assistant activity
Login Users with and without Authenticator 2FAHow far two-factor authentication has been taken up
2FA Users with Fewer than 3 Backup CodesUsers at risk of locking themselves out

Two rows are worth acting on rather than just reading:

  • 2FA Users with Fewer than 3 Backup Codes tells you who is heading for a lockout. A user who runs out of backup codes and loses their authenticator needs account access assistance to get back in, so it is cheaper to have them reset their backup codes first.
  • Active Elevated Users is worth a periodic glance against who actually needs elevated access.

A rise in Failed Login Events is usually someone struggling with a new password, but a large jump is worth investigating.

The same view and edit counts as above, but for automated access, split so you can tell the two apart:

  • API Users, API Views and API Edits for the REST API
  • MCP Users, MCP Views and MCP Edits for AI assistants

Keeping these separate from the interface figures answers “how much of our activity is now automated”, and makes an unexpected assistant or integration visible.

Host-level figures for platforms on their own dedicated server: server uptime, logical CPUs, load average, memory available, filesystem free, and the PHP version, memory limit, execution time and upload size limits the platform is running under. Usage bars show server memory and platform filesystem use as percentages.

Filesystem Free is the one that causes trouble if ignored: billing runs generate bills, CDR files and disaster recovery archives, and a platform that runs out of disk stops doing all three.

Two caches that keep the platform quick. Each shows whether it is enabled, its hit rate and memory use, and how much it is holding.

A healthy hit rate is high. A low hit rate, or a cache reporting itself as full or heavily fragmented, is worth raising with support rather than acting on yourself.

Panels degrade rather than error. A figure the platform cannot read shows as Unavailable, and a panel that cannot be built says so, leaving the rest of the page working.

If the queue is backing up, disk is filling, or the security counts look wrong, contact support with what the page shows.