The phone that was passed on Friday
The on-call mobile is handed over on Friday afternoon. This week the engineer forgot to take it home, so the out-of-hours calls diverted to a colleague who was not expecting them. A client's server went down at eleven at night; the RMM alert went to an email inbox nobody watches out of hours. The client called the main number, which went to voicemail.
In the morning, the service desk finds a voicemail, an alert and a very unhappy client. Meanwhile, another client with no after-hours cover in their contract rang the on-call engineer at nine in the evening about a printer, and the engineer, not knowing their contract, spent an hour fixing it.
Why on-call cover breaks down
On-call is run informally because it is only a small part of the week, but it is where service failures are most visible.
- Rotas are in a spreadsheet, and swaps are agreed in chat.
- Phone diversion depends on someone changing a setting.
- Critical RMM alerts go to email, not to a person.
- There is no escalation if the on-call engineer does not answer.
- Engineers do not know which clients have after-hours cover.
Contracts add another layer. Some clients pay for full after-hours cover, some for critical systems only, some for none. The engineer answering at night rarely has that information to hand, so they either help everyone, giving away evenings, or hesitate with a client who is entitled to help.
What weak on-call costs
| Failure | Consequence |
|---|---|
| Call reaches nobody | A client with cover gets no answer |
| Alert unseen overnight | An outage discovered by the client in the morning |
| Uncovered client helped for free | Engineer evenings given away |
| No logging | Morning handover by memory, and unbilled after-hours work |
| Burnout | Engineers dread their on-call week |
After-hours cover levels come from your contracts. The setup applies them and tells the engineer what each caller is entitled to.
The on-call setup we build
- An on-call rota in a shared calendar or scheduling tool, with swaps made there and reflected everywhere automatically.
- After-hours calls to your support number routed by your phone system (through its API, or a service such as Twilio) to the current on-call engineer.
- Critical RMM alerts for clients with after-hours cover sent as calls or push notifications to the on-call engineer, with acknowledgement required.
- Escalation to a second engineer and then a manager if nobody acknowledges within a time you set.
- Caller lookup against the PSA, showing the engineer the client, their contract's after-hours cover and any open tickets, before they answer.
- Every call and alert logged to the PSA as a ticket with times, so the morning handover is written and after-hours work can be billed where the contract allows.
An on-call week that works
The rota is in the calendar, and the engineer on call gets the calls and critical alerts on their own phone. At eleven at night, the server alert for a covered client rings the engineer, who acknowledges and starts work. If they had not acknowledged, the second engineer would have been called. The client who rings about a printer is shown as having no after-hours cover, so the engineer logs it for the morning and tells the client politely.
In the morning, the service desk reads the overnight tickets rather than listening to voicemails. At month end, after-hours work for clients whose contracts allow billing is on the invoice. Engineers see a fair rota and know exactly what is expected on their week.
Over a few months, the logs show what after-hours cover really involves: how many calls and alerts arrive, at what times, for which clients, and how often the second engineer had to be called. That is the evidence you need to decide whether on-call is fairly paid, whether a client with frequent night-time alerts needs a fix rather than more cover, and whether your after-hours tier is priced to match the work it brings.
Is your on-call cover fragile?
- On-call means a shared mobile passed between engineers.
- Critical alerts out of hours go to email.
- Nobody is called if the on-call engineer misses a call.
- Engineers do not know which clients have after-hours cover.
- After-hours work is not logged until the next day, if at all.