PagerDuty Contractual response time Shared senior rota

24/7 IT on-call support —
your servers watched, even at 3am

Proactive monitoring, PagerDuty alerting, a guaranteed 4h or 1h response time, 365 days a year. Your team sleeps, we watch.

Your server goes down at 10pm. Who picks it up?

Most critical failures happen outside business hours. Without a structured rota, every minute of downtime costs money — and nobody sleeps well.

Your admins are exhausted

Your sysadmins are on call on top of their day job. The result is burnout, turnover, and mistakes made at 3am on a machine nobody should be touching half-asleep.

Out-of-hours failures cost twice as much

No team, no process, unpredictable reaction time. Every hour of downtime is revenue lost. Work out the impact.

French labour law constrains you

Allowances, compensatory rest, a collective agreement, travel time. If you employ people in France, an in-house rota is a legal and HR problem before it is a technical one.

A shared rota, structured and transparent

Our on-call service is not a phone number. It is a complete system: proactive monitoring, automatic escalation, documented intervention.

Proactive 24/7/365 monitoring

Agents on every server: CPU, RAM, disk fill level, services, TLS certificate expiry. We catch problems before the outage, not after it. On on-call only we plug into your monitoring; with managed services we deploy ours.

PagerDuty call-out

On a confirmed production outage you trigger an incident through PagerDuty and our rota is paged immediately. Triggering outside a real outage is billed at €100 excl. VAT.

Contractual response time: 4h or 1h

No best effort. The response time is contractual and measurable — 4h or 1h depending on the level you choose, 24/7/365. We do not commit to a resolution time, and we explain why below.

A shared senior rota

No junior alone at night. Our engineers are seniors with 10+ years on Linux, networking and virtualisation. There is no level 1 here: your incident reaches an engineer directly.

Monthly report

Every month: alerts raised, incidents handled, time spent, measured availability, preventive actions. Grafana dashboards stay available in real time between two reports.

Support: email and member portal

Routine support goes through email or your member portal. No phone line. PagerDuty is reserved for confirmed production outages — that is what keeps it answered.

From audit to live cover in one week

No three-month project. Onboarding is fast and structured.

1

Entry audit

Thirty minutes to map your infrastructure, identify the risks and size the cover you actually need.

2

Runbooks

We write the intervention procedures specific to your infrastructure: critical services, contacts, escalation paths.

3

Monitoring and PagerDuty

Your monitoring wired to PagerDuty, or ours deployed with managed services; alert thresholds set, escalation configured and tested.

4

Cover goes live

A date and a time. From that moment the alert wakes us and the response time runs. Monthly reporting, runbooks improved as we go.

Outsourced on-call is for you if…

SMB / mid-market

You run critical servers but have no 24/7 IT team

Your ERP, your storefront or your APIs cannot wait until Monday morning. But a compliant 24/7 rota needs four employees: about €4,500 a month in on-call allowances if they are already on staff, €34,800 a month if you have to hire them. Outsourced, 24/7 on-call starts at €90 per server per month, €405 for 5 servers with a 4h response time; the €3,450 fleet package only pays off at around fifty servers.

E-commerce / SaaS

Continuous availability is part of the product

Your customers buy at any hour. One hour of downtime on a Saturday evening means lost orders and customers who do not come back. You need a contractual response time, not a mobile number ringing into the void.

IT teams

Your team wants to hand over nights and weekends

Your admins handle the day perfectly well. It is the nights that wear them down and drive turnover. Hand over the 24/7 standby and let your team spend its daylight on projects.

Multi-host

Infrastructure spread across several providers

Your servers sit across OVHcloud, Scaleway, Hetzner, Contabo or IONOS. One on-call contract covers the lot, whoever hosts what.

What we commit to, and what we refuse to commit to

We commit contractually to a response time (GTI): 4h or 1h depending on the level you choose, 24/7/365. It is a delay we control, because it depends only on how our rota is organised.

We do not commit to a resolution time (GTR). Until the diagnosis is made, nobody can honestly promise a repair duration: a service that refuses to start may take two minutes or turn out to be data corruption. A resolution time printed on a brochure is either a penalty provisioned into the price, or a promise that gets renegotiated on the day of the outage.

What we give you instead is a measured figure rather than a guaranteed one: the restore duration recorded during the last test. A telecom operator can commit to a restoration time on a link because the failure modes there are bounded and known — a cut fibre, a terminal to swap. A system incident has no such property. The distinction is set out in our glossary entry on operator GTR.

On-call plans

24/7/365 cover, billed per server. The same price whoever hosts your machine — it is engineering time, not a share of what the hardware costs. Two ways to buy it: on its own, wired to your monitoring, or included in a managed server contract that also carries monitoring, security updates and operational maintenance.

4h response time
24/7/365
1h response time
24/7/365
On-call only
You provide the monitoring: it is what triggers the call-out. PagerDuty alerting, intervention, written report.
€90 excl. VAT/month* €160 excl. VAT/month*
Managed services, on-call included
We also carry monitoring, security updates and operational maintenance, with 1h of intervention a month included, then €120 excl. VAT per hour.
€150 excl. VAT/month*
Pro plan
€250 excl. VAT/month*
Critical plan

* Excludes the server / VM you rent separately. On-call only gets a volume discount from 5 servers; beyond roughly fifty, the fleet package (€3,450 or €6,000 per month) works out cheaper. On a 7am-10pm, 7-day window, on-call is sold as a fleet package (€1,800 with a 4h response time, €3,000 with 1h); the Essential managed plan (€70) covers the same window, without nights. You subscribe to the machine directly with the host and the contract stays in your name: we only bill the management service. Request a tailored quote →

In-house versus outsourced: the real cost

The cost of an in-house rota goes well beyond the on-call allowance. Full cost breakdown.

Criterion In-house rota RDEM — outsourced
Annual cost ~€418,000 (four staff: loaded payroll plus allowances) €1,080 per server (on-call only, 4h response time)
Minimum headcount Four employees in rotation, compensatory rest included A pooled rota of independent senior engineers
French labour law Compensatory rest, collective agreement, allowances Nothing to manage — it is our obligation, not yours
Availability Variable — holidays, sick leave, resignations 365 days a year, contractually
Tooling Yours to buy and run Monitoring and PagerDuty included

Need more than on-call?

Our managed server contracts include 24/7 on-call plus preventive maintenance, proactive updates and off-site NimbusBackup PBS backups — one contract per server, €150 (4h) or €250 (1h) per month for 24/7 cover.

And when the outage itself is the problem — not just the response time — the next step is architectural rather than contractual. In a two-layer architecture, the service fails over on its own within seconds and the redundancy rebuilds itself; on-call then only steps in cold, to check what is left to do.

Frequently asked questions

What happens if a server fails at night?

PagerDuty alerts our on-call rota immediately. A senior engineer takes ownership of the incident, diagnoses it and intervenes within the contractual response time (4h or 1h depending on your plan). You receive a written report the following morning.

What response time do you commit to?

A 4h or 1h response time (GTI), 24/7/365, depending on the level you choose — on-call only, or included in the Pro (4h) and Critical (1h) managed plans. We commit to a response time, not to a resolution time: until the diagnosis is made, nobody can honestly promise a repair duration.

How does escalation work?

PagerDuty pages the on-call engineer. If the incident needs a level 2 or level 3 skill — networking, databases, security — escalation is automatic. Runbooks specific to your infrastructure are written during onboarding to shorten the diagnosis.

How much does outsourced on-call support cost?

€90 excl. VAT per month per server with a 4h response time, €160 with 1h, 24/7, with a volume discount from 5 servers. You provide the monitoring: it is what triggers the call-out. Beyond roughly fifty servers, the fleet package (€3,450 or €6,000 per month) works out cheaper. If you want us to carry monitoring, security updates and operational maintenance as well, on-call is included in managed services: €150 per month with a 4h response time (Pro), €250 with 1h (Critical). Prices exclude the server or VM you rent separately from your host.

Which technologies do you monitor?

Linux (Debian, Ubuntu, RHEL, Rocky), Proxmox VE, Docker, PostgreSQL, MariaDB/MySQL, Nginx, Apache, HAProxy, PHP-FPM, Node.js, Redis, Elasticsearch. We also monitor network components, TLS certificate expiry and the business-specific services documented in your runbooks.

Can we outsource nights and weekends only?

Not as a separate tier: a standby that stops at 8am does not cover an incident at 7:30, so to cover nights and weekends, on-call is bought as 24/7, from €90 per server per month. In practice, your team keeps the day, the projects and user support; we only take the incidents it does not handle.

How do you handle unjustified call-outs?

A PagerDuty trigger outside a confirmed production outage is billed at €100 excl. VAT per incident. That rule protects on-call availability for genuine emergencies. Routine support — questions, planned changes — goes through email or the member portal.

Do you cover infrastructure hosted outside France?

Yes. On-call work is remote by nature, so the location of your servers does not change what we commit to. We routinely operate infrastructure at OVHcloud, Scaleway, Hetzner, Contabo and IONOS across their European datacenters. What matters is that we can reach the machine and that you hold the account with your host — the contract stays in your name, we never resell infrastructure.

Why does French labour law matter when buying on-call?

Because it is what makes an in-house rota expensive in France. As soon as an employee is placed on call, French labour law imposes a strict frame: rest periods, compensation, and a collective or company agreement. Running a sustainable 24/7 rota needs four employees in rotation. Outsourcing removes that entirely — you buy a service, not working time.

What does an in-house 24/7 rota cost?

A sustainable 24/7 rota needs four full-time employees, not four allowances paid on top of a day job: compensatory rest and the 11-hour daily rest consume working days, and someone woken at 3am does not work the next day. Count €31,600 per month in employer cost for four people, plus about €3,200 per month in allowances — €34,800 per month, close to €418,000 a year. Outsourced on-call costs €1,080 a year per server with a 4h response time, €1,920 with 1h.

Who staffs your on-call, if a rota needs four FTE?

The four-FTE rule applies to a salaried team: the 11-hour daily rest and compensatory rest required by labour law consume working days. Our rota is staffed by independent senior engineers, who fall outside that framework, and pooled across several clients. That does not make fatigue go away: which is why every alert comes with its runbook, escalation to the second on call is automatic, and the rota is growing — the aim is to run it with four to six people, employees and partners included.

Your servers deserve a structured rota

A 30-minute audit of your infrastructure. We identify the risks, you decide what to do about them.