Skip to content
plugboard.appRequest a demo

Running the desk

Services

/services, and /services/<id> for one. Needs service.view, with the Tickets module on.

A service is something the school runs and people rely on: Wi-Fi, Compass, Printing, Student laptops, Email. Each one has an owner, support hours, the campuses that use it, the other names staff call it, a criticality and a review date. It lists the records it depends on and the contracts and suppliers behind it, and it says whether it is working, in plain words.

Services tie together records the desk already keeps. A monitor, a network device, a device or a printer shows the services that depend on it under Used by, a ticket is filed under the service it names and measured against the service’s own service levels, a change names the services it affects, and the school’s public status page can list the services people use.

Services, under Service desk next to Changes, lists one row a service:

ColumnWhat it shows
ServiceIts name, with a line under it when there is something to say: High criticality or Low criticality, its open tickets, Review overdue, a service it depends on that is in trouble, such as Wi-Fi is having problems, which Compass depends on, or a provider service with a problem, such as Exchange Online has a problem, which Email depends on
StatusWorking, Having problems, Down, Planned work or Not checked, with Set by a person while a person’s status lasts
OwnerWho looks after it, or No owner
CampusesThe campuses that use it, or Whole school

Services for IT only, such as RADIUS or the print server, are listed under Supporting services, below the ones people use. They are kept off the status page.

Search covers names, other names and owners. Add service is for people with service.manage. While the school has no services, somebody who manages services and can see monitors is offered Start from a monitor.

Add service asks for:

Field
NameWhat the school calls it, such as Compass. Each name is used once
Who uses itStaff and students, or IT only (a supporting service)
Other namesWhat staff call it, separated by commas, such as the portal, roll marking. A ticket that names one is filed under this service. Up to 20
DescriptionWhat it is, in a sentence or two
OwnerWho looks after it, or No owner
CriticalityHigh, Normal or Low
Support hoursThe desk’s hours, School days or Weekdays, all year with a start and end time, or Around the clock. School days are closed between terms. The service’s tickets are measured in these hours; see service levels
Review everyMonth, 3 months, 6 months or 12 months, or Never. A service starts on 3 months
CampusesShown when the school has campuses. None ticked is the whole school
Show on the status pageOffered for a service staff and students use

A school can have up to 200 services. Edit on a service changes any of it, and Delete service…, in the menu beside Edit, removes it: what it depends on is unlinked, its service levels go with it, its documentation is unbound, and changes and major incidents that named it name it no longer. Its tickets stay, losing only the link to it.

A school that already watches its systems with monitors can start from them. Make a service from this monitor, on a monitor’s detail, or Start from a monitor on an empty Services page, makes a service with the monitor’s name and other names, depending on that monitor, with the monitor counting towards its status. Then fill in the owner, hours and the rest with Edit. It needs service.manage and monitor.view.

The page reads from top to bottom:

  1. The description.
  2. The status: the status in words, since when, a line naming any service it depends on that is down or having problems or any provider service with a problem, the change or major incident behind it (see changes and major incidents), and a person’s message with Set by a person until and the time it ends. People with service.manage have Set status, or Change or renew and Clear while a person’s status lasts.
  3. Changed in the last 72 hours: the changes that name the service or something it depends on. See what did we change?
  4. Depends on, grouped into Services, Monitors, Network devices, Devices, Printers and Provider services, with the Fleet at the end. Each part links to its own page and shows its own state, such as Up, Down or Ready, and a monitor says whether it Counts towards its status. Add a part links another.
  5. Supported by: the software contracts and suppliers behind it.
  6. Used by: the services that depend on this one.
  7. Service levels: its support hours and its targets for each priority, while service levels are on. See service levels.
  8. Open tickets, each with the words it matched on when a match filed it, and Earlier tickets folded away under them.
  9. Documentation bound to the service.

The column beside it holds Owner, Support hours while service levels (module.sla) are off, Campuses, Other names, Criticality, Who uses it, Status page (Shown or Not shown), the Review: the next review or Overdue since, how often, when it was last reviewed, and Mark reviewed; and Service review, whose Open runs the Service review report for this service alone, for people who can open Reports.

Add a part asks for the kind of record, then finds it by name:

KindWhat it isSeen with
ServiceAnother service, such as RADIUS under Wi-Fiservice.view
MonitorA monitor, which can count towards the statusmonitor.view
Network deviceA switch, access point or gateway from Networknetwork.view
DeviceA device from the device registerdevice.view
PrinterA printerdevice.view
Provider serviceA Microsoft 365 or Google Workspace service, such as Exchange Online. See provider servicesmonitor.view
Contract (supports it)A software contractsoftware.view
Supplier (supports it)A supplierstock.view

The picker offers only the kinds you can see. On the page, each part shows only to people who may open that record. A part you cannot open is counted, never named: And 3 more you do not have access to.

A service cannot depend on itself, either directly or through another service: linking one that would make a loop is refused. The ⋯ beside a part removes it, and on a monitor, Count towards its status or Stop counting towards its status changes whether it decides the service’s status. A service can have up to 1,000 parts. A whole kind of device, such as every student laptop, is a fleet rather than a part each.

Status does not pass from one service to another. When Wi-Fi is having problems, Compass keeps its own status and says Wi-Fi is having problems, which Compass depends on, so the desk sees the likely cause without Compass reading as broken when its own checks pass.

A service’s status is one of five, in plain words:

StatusMeans
WorkingEvery monitor that counts towards it is up, or a person says so
Having problemsAt least one monitor that counts is down, or a person says so
DownEvery monitor that counts is down, an open major incident names it, or a person says so
Planned workA change that names it is under way inside its window
Not checkedNo monitor counts towards it, and nothing else says

The first of these that applies decides it:

  1. A person’s status, while it lasts.
  2. An open major incident that names the service: Down.
  3. A change that names the service, in progress inside its window: Planned work.
  4. The monitors that count.

So a monitor that goes down during planned work leaves the service reading Planned work, and a status a person sets wins over a major incident until it ends.

A monitor counts when it is linked with Counts towards its status, is not paused and has a result. A monitor that has not been checked yet is no evidence either way. Status from monitors needs the Monitors module; with it off, a service reads Not checked unless a person, a major incident or planned work says otherwise.

Monitors do not see everything. When Microsoft says mail is running late, or the library system is out for the afternoon, Set status says so:

  • Status: Down, Having problems or Working.
  • Message: up to 280 characters. On a service shown on the status page it is public too, so say what people should do, not what broke.
  • For: 1, 2, 4, 8, 12 or 24 hours.

A person’s status wins over everything else until it ends. Change or renew sets it again, starting the hours afresh, and Clear hands the service back. The desk checks every minute, so a status that has run out goes back to the monitors without anybody touching it.

Service levels on a service’s page holds its support hours and its first response and resolve targets: a row for each priority, and a row for each record type the service or the desk sets on its own. Where the service sets none, the desk’s targets apply and are shown in grey, under Grey targets are the desk’s, where this service sets none, or Compass uses the desk’s targets when it sets none at all. A target nobody sets reads None.

Anybody who can see the service can read its service levels. The section is left off while service levels (module.sla) are off, and the support hours show in the column beside the page instead.

Edit, for people with service.manage, opens Service levels for Compass:

  • Support hours: the same choices as the service’s own form.
  • Targets: a First response and a Resolved within for each priority, in minutes, hours or days. Leave one empty to use the desk’s: the empty box says what applies instead, such as Desk: 4 hours, or Any type: 2 hours for a type’s row.
  • Add targets for this type gives a priority a row of its own for one record type: Incident, Service request, Problem or Change.

Save service levels saves the hours and the targets together. A campus manager changes them only on a service whose campuses are all theirs.

A ticket filed under a service looks for its target in this order, first response and resolve each on their own:

  1. The service’s target for the ticket’s priority and type.
  2. The service’s target for the priority, any type.
  3. The desk’s target for the priority and type.
  4. The desk’s target for the priority, any type.

A service that sets only a tighter resolve time keeps the desk’s first response time. A ticket with no resolve time at any step has no target, as on the desk. A ticket filed under no service uses the desk’s two steps.

A ticket filed under a service with support hours of its own is measured in those hours. A Compass ticket raised at 4:30 pm, when Compass is supported on school days from 8 am to 4 pm, starts counting at 8 am on the next school day. A service on The desk’s hours uses the desk’s working calendar. The ticket queue, the Service levels page, overdue alerts and reports all measure a ticket the same way.

A ticket that a match files under a service when it is raised is dated against that service’s targets and hours from the start. Choosing a service for an open ticket, choosing another or clearing it, works its due dates out again from when it was raised, as a priority change does: a five-hour-old ticket given a service with a four-hour target is already late. A paused ticket is measured as of its pause, and a ticket already answered keeps its first response date. Time the ticket has already spent waiting moves onto the new service’s hours as the same share of its life, so a ticket that waited half of its four days on a desk open around the clock has waited half of the service’s hours in those days, not two full days of them. Resolved and closed tickets keep the dates they were measured against.

Changing a service’s targets or support hours does not re-date its open tickets, as changing the desk’s does not. Tickets filed under it afterwards use the ones in place then.

A change names the services it affects, and a major incident can name the service it is about. Both decide what the service reads while they last.

While a change that names a service is In progress inside its window, the service reads Planned work on Services, on its page, on Home and on the public status page. Its page says which change, such as Planned work: Upgrade the Wi-Fi controller firmware, until 7:00 am, to people with change.view. Planned work ends when the window does, or when the change moves on, whichever comes first, and the service goes back to its monitors.

Planned work is neither an outage nor a recovery, so starting or ending it starts no flow that waits for a service to go down or come back.

While an open major incident names a service, the service reads Down, and its page links to it, such as Major incident: #1042 Wi-Fi down in the hall, since 20 minutes ago. A major incident whose ticket you cannot open is counted, never named. When the major incident ends, or its ticket is resolved, the service goes back to its monitors.

Changed in the last 72 hours, under a service’s status, answers the first question when a service goes wrong. It lists the changes in progress, or implemented with a window in the 72 hours before the service went down or started having problems, or in the last 72 hours while it is working, that:

  • name the service (Names this service); or
  • link something it depends on: a service, monitor, network device or device, or a device or network device in its fleet, such as Linked to AP-B12, which Wi-Fi depends on.

Each shows its status and window, and those that name the service come first. Only records you can open explain a change, so a change linked only to a record you cannot see is left out. A change already shown as planned work is not listed again. It needs change.view, and says nothing when no change is near. It is worked out by rules from the records, with no AI.

Some services rest on one Microsoft or Google runs: Email on Exchange Online, video calls on Google Meet. Add a part, Provider service, finds these by name, listed as Exchange Online (Microsoft 365). The picker offers the services named in the incidents Plugboard has read from the provider: Google Workspace’s public status dashboard, and Microsoft 365 service health for the school’s own tenant. A provider service is there to choose once the provider has reported on it.

While the provider has an open incident on it, the part reads Has a problem, and the service’s page, the Services list and Home say Exchange Online has a problem, which Email depends on beside the service’s own status. It never changes the status: Email still reads Working while its own monitors pass, because a provider’s incident often touches some schools and not others.

Without an open incident the part shows no state, because a provider saying nothing is not proof all is well. An advisory, a resolved incident and a provider status that has not been read lately say nothing either. The part links to the provider’s own status page.

Provider services need monitor.view and the Integrations module. Each reader sees a provider’s problem only if they may read that provider’s status, and somebody without monitor.view sees the part counted, never named.

A service used across a whole kind of device, such as Student laptops, covers them by rule instead of one link per device. Set a fleet, at the end of Depends on, opens Fleet for Student laptops:

  • Devices: Chromebooks, Laptops, iPads, Desktops, Tablets, iPhones, Phones, Printers and Other devices, the device types in the device register.
  • Network devices: Access points, Switches, Gateways, Firewalls, Controllers and Other network devices.
  • Campuses, when the school has them. None ticked is every campus.

Tick at least one kind of device or network device. The dialog reads the rule back, such as Chromebooks and laptops at Senior campus or Access points across the school, and Save fleet keeps it.

Fleet then shows the rule and how many records it matches, such as 1,412 devices and 38 more you do not have access to. Records on campuses you do not work at, or of a kind you cannot see, are counted, never named. A rule over one device type and no campus links its count to the device list filtered to that type. A rule that names campuses leaves out records with no campus. Change the fleet… and Clear the fleet are in the menu beside it.

The rule is matched each time it is read, so a laptop added to the register is in the fleet without anybody linking it. A fleet never counts towards the service’s status. It shows under Used by on each device and network device it takes in, and in what did we change? and the services a change is offered.

Setting a fleet needs service.manage. Somebody limited to some campuses names only their own campuses, and cannot leave the campuses empty: a fleet across the whole school needs somebody who looks after the whole school.

A monitor’s detail, a network device’s expanded row, a device’s side column and a printer’s expanded row show Used by: the services that depend on that record, with their status. A monitor also says whether it counts towards each one’s status. A device or network device also lists the services whose fleet takes it in, marked by rule; one linked directly and by rule is listed once, as linked.

Nothing is shown when no service depends on the record, or when you cannot see services. A monitor’s detail also offers Make a service from this monitor to people with service.manage.

When a ticket is raised, Plugboard looks for the services it names: a service’s name or one of its other names in the subject or the body, or a monitor the service depends on. It matches names of three characters or more, by rules, with no AI. A service used only at other campuses than the ticket’s is never offered.

The ticket namesWhat happens
Exactly one serviceIt is filed under that service, and the Service row in the ticket’s details reads Matched on and the words, such as Matched on “Wi-Fi”
Several servicesEach is listed under Suggested, with Use, and Choose another opens the school’s services
NoneThe Service row reads No service

People with ticket.update choose the service with Change, or take it off with Clear. On an open ticket that works out its due dates again; see when a ticket’s service changes. A person’s choice is never overwritten, because matching runs only when the ticket is raised. A ticket that carries on from another keeps its service. Requesters are never asked to choose one on the portal.

Filing a ticket under a service is recorded in the audit log as Ticket filed under a service. It does not count as working the ticket, so filing a settled ticket under a service leaves its approvals as they were.

A document can be bound to a service under Shown on, with the Service kind. It shows on the service’s page and in the documentation on every ticket filed under the service, ahead of a runbook for a monitor the ticket happens to name. Binding one needs service.view as well as permission to edit the document.

Home. The Service health widget lists services first, most critical first, each with a service it depends on that is in trouble, a provider service with a problem, the person’s message, or its monitors underneath. Its verdict says Email is having problems, 2 of 6 having problems or All 6 working. Monitors that belong to no service and are not up follow as one row. A high-criticality service that is down replaces the verdict with a band across the widget. Without any services the widget is the monitor board.

Needs you. A service’s owner sees A service you own is due for review, naming the service and when it fell due, such as RADIUS, overdue since 1 Oct.

Reports. The Service review reports each service’s tickets, downtime, targets met, changes and problems for a period. SLA attainment and Open problems group by service, Issue insights shows issues by service, and the tickets dataset has a Service field, with No service for tickets filed under none, so a board can count tickets by service.

The public status page. Services marked Show on the status page are listed there with their status, Planned work included, the person’s message while it lasts, and 90 days of uptime where a monitor counts towards them.

A flow can start when a service goes down or starts having problems, when it is working again, and when its review falls due. A webhook can be sent service.down, service.up and service.status. Planned work starting or ending is a service.status only.

Flows never change a service: its status, parts, service levels, fleet and review are kept by people, and so is the service a ticket is filed under.

Each service has a review date, three months from when it is added to begin with. The owner sees a due review in Needs you, the list says Review overdue, and an overdue service’s page leads with Review overdue since and Mark reviewed for people with service.manage.

Mark reviewed records who reviewed it and when, and sets the next date from Review every. A service set to Never is Not on a schedule. A review is a good time to check the owner, the other names people use, its service levels, and that what it depends on is still right. A service review with school leadership is a natural place to do it, with the Service review report in hand.

A service used at certain campuses is seen by people at those campuses, and a whole-school service by everyone who can see services. Somebody limited to some campuses changes only services whose campuses are all inside theirs; a whole-school service needs somebody who looks after the whole school, and a refused change says so.

The most an AI model is given about a service is its name and status: Describe a widget on a report board can offer service names as values of the tickets data’s Service field. What a service depends on, its fleet and its service levels never reach a model, and Ask cannot run the Service review.

Each change is recorded in the audit log, with the record type Service:

Shown asWhen
Service addedAdded by hand or from a monitor
Service updatedEdited, including its name; its service levels (changed the service levels of Compass); or its fleet (set the fleet of the service Student laptops, or cleared the fleet)
Service deletedDeleted
Part linked to a serviceA part linked, or a monitor starting or stopping counting towards its status
Part removed from a serviceA part removed
Service status setA person’s status set, with its message, or cleared
Service reviewedMark reviewed
Ticket filed under a serviceA person chose or cleared a ticket’s service

A change naming a service is recorded on the change, such as said a change affects Wi-Fi, and naming the service on a major incident as Service named on a major incident, on its ticket. Status worked out from monitors, changes and major incidents, and a person’s status running out, are not recorded.

What you seeWhy
A service reads Not checked with monitors linkedNone of them counts towards its status, they are paused or not yet checked, or the Monitors module is off
A service stays Down after its monitors recoverA person’s status is set, or an open major incident names it. Clear hands back a person’s status; ending the major incident, or Change service… on it, hands back the other
A service reads Planned work while a monitor is downA change that names it is in progress inside its window, and planned work comes before the monitors. It ends with the window
Planned work did not show during a changeThe change is not In progress, its window has not started or has ended, or it does not name the service under Services it affects
And 2 more you do not have access toParts you cannot open are counted, never named. Ask for the record’s view permission if you need them
A ticket was not filed under a service it namesIt named several, so they are suggested; the name is shorter than three characters; the service is used at another campus; or somebody chose for it
A ticket’s due date moved when its service changedThe other service has other targets or support hours, so the dates were worked out again from when it was raised
Open tickets kept their due dates after a service’s targets changedChanging targets or hours dates only tickets filed under it afterwards
Service levels is missing from a serviceService levels (module.sla) are off. The support hours are in the column beside the page
A provider service is not in the pickerThe provider has not reported an incident naming it, you do not have monitor.view, or the Integrations module is off
A device is not in a fleet it should be inIts device type or campus is not the rule’s, or the rule names campuses and the device has none
Choose your campuses. A rule across the whole school needs someone who looks after the whole school.Your access is limited to some campuses, so a fleet must name only yours
This service is used beyond your campusesYour access is limited to some campuses. Somebody who looks after the whole school can change it
That would be a loopThe other service already depends on this one, directly or through another
Show on the status page is missingThe service is for IT only. Supporting services never go on the status page
PermissionAllows
service.viewSee the Services page, a service and its service levels, Home’s services and Used by, each part with that record’s own view permission
service.manageAdd, edit and delete services, link and remove parts, set a fleet, set service levels and support hours, set or clear a status, mark reviewed, and make a service from a monitor
ticket.updateChoose or clear a ticket’s service
change.manage with service.viewName the services a change affects
ticket.create and ticket.update with service.viewName, change or clear the service a major incident is about
report.view with service.viewRun, export and be sent the Service review

service.view is free, and every built-in role that can see tickets holds it. service.manage makes an account a Business seat, and only the Owner role holds it to begin with. Give it to the roles that look after services under Settings, Roles. See permissions.