System Availability Rate
Definition
System Availability Rate
System availability rate is the share of an agreed period during which a system is up and usable by the people depending on it. It is usable time divided by total time, judged from the user side rather than the server room.
Usable is the operative word — a platform can be running perfectly while a broken login page keeps everyone out.
The measure drives contract terms as much as engineering. Most outsourcing agreements express it as a target with credits attached to a miss.
Where you measure it matters enormously. Availability at the network edge and availability at the application often differ by a visible margin every single month.
Exclusions do the rest of the work — agreements that carve out all planned maintenance produce flattering numbers and frustrated users.
Key takeaways
- System availability rate divides usable time by total time in an agreed period.
- Measurement point and exclusions move the figure more than engineering does.
- Each additional nine cuts allowed downtime by roughly ninety percent.
- Availability should be measured from the user side, not from the infrastructure.
How it works
Define the measurement window, agree what counts as available, subtract unavailable minutes, then divide by total minutes. Express the result as a percentage or in nines, and state whether planned maintenance sits inside or outside the calculation.
Targets are easier to read as allowed downtime. Nines look similar on paper and translate into very different operational commitments.
| Target | Allowed downtime per year | Allowed downtime per month |
|---|---|---|
| 99% | About 3 days 15 hours | About 7 hours 18 minutes |
| 99.9% | About 8 hours 46 minutes | About 43 minutes |
| 99.95% | About 4 hours 23 minutes | About 22 minutes |
| 99.99% | About 53 minutes | About 4 minutes |
Contingency planning underpins the target. NIST Special Publication 800-34 sets out how organisations should evaluate systems and operations to determine contingency planning requirements and priorities.
Security incidents belong in the same conversation. The Cybersecurity and Infrastructure Security Agency frames operational resilience as part of cyber hygiene rather than as a separate engineering concern.
Availability also needs a severity view. A system that is technically up while its slowest function times out is unavailable to the people trying to use that function.
Recovery speed belongs beside the percentage. Two providers can report identical availability while one restores service in eight minutes and the other takes two hours to notice anything broke.
Agree the reporting evidence early. Monitoring exports both parties can query settle far more disputes than a monthly summary produced by one side alone.
Examples
Availability disputes almost always trace back to where the measurement sits and what the agreement excludes. Four cases show how the same infrastructure produces different numbers.
A Philippine shared services platform. Edge monitoring reported 99.96% while application probes reported 99.71%. The gap was a slow authentication service nobody was watching.
A hosted contact centre. Planned maintenance ran to nine hours a month and was fully excluded. Users experienced 98.7% and the contract reported 99.95%.
A payments integration. The provider met 99.9% across the year but concentrated its downtime into two long outages, which cost the client far more than a hundred short ones would have.
A regional logistics system. Adding a second data centre moved availability from 99.5% to 99.97% — a change in architecture rather than in effort.
Related terms
Availability connects the contract that sets the target with the design and monitoring work that meets it. The terms below cover the agreement, the fallback, and the watchers.
- Service Level Agreement (SLA): the contract that carries the availability target.
- Business Continuity Plan (BCP): the arrangement invoked when availability fails badly.
- Live Monitoring: the detection layer that notices an outage first.
- Escalation Plan: the route an outage follows once detected.
- Global Delivery Center: the multi site model that spreads availability risk.
- IT Support Technician: the role restoring service at the front line.
- Key Performance Indicator (KPI): the reporting family the measure belongs to.
FAQ
What is a good system availability rate?
99.9% suits most business systems, and 99.99% suits payments or safety critical work. Higher targets cost far more than they usually return.
How much downtime does 99.9% allow?
About 43 minutes a month, or roughly 8 hours 46 minutes a year. That is the whole budget, planned and unplanned together.
Should planned maintenance be excluded?
Only within a narrow agreed window, and it should always be reported. Blanket exclusion makes the figure meaningless to users.
Where should availability be measured?
As close to the user as possible. Infrastructure level probes consistently report better numbers than users experience.
Is availability the same as uptime?
Closely related but not identical. Uptime asks whether the system is running; availability asks whether it is usable.
Why do two parties report different figures?
Because they are measuring at different points or applying different exclusions. Agree both in the contract before the first report.
Find delivery partners and technical capacity in the Outsource Accelerator BPO hubs.
Independent
