Uptime Availability Calculator

Calculator

Choose a method, set your measurement period, and include the reliability options you track.

Calculation method

Pick what matches your monitoring and engineering data.

Measurement period

Used for availability and SLA error budgets.

SLA target (%)

Example targets: 99.0, 99.9, 99.95, 99.99, 99.999.

Custom duration (days)

Custom duration (hours)

Custom duration (minutes)

Start (local)

End (local)

Planned maintenance (minutes)

Optional maintenance time in the same period.

Exclude planned maintenance

Removes planned time from the denominator window.

Service-level downtime

Use measured downtime during the selected window. Add an impact factor for partial outages.

Unplanned downtime (minutes)

Total unplanned downtime minutes in the window.

Impact weighting (%)

100% for full outage, 50% for half capacity.

Quick check

If you measured downtime at the user edge:

Use impact = 100% and exclude planned if your SLA does.

Reliability (MTBF/MTTR)

Model availability from engineering reliability metrics, then estimate expected downtime for your period window.

MTBF (hours)

Mean time between failures.

MTTD (hours)

Mean time to detect (optional).

MTTR (hours)

Mean time to repair/restore.

n (components)

Total identical components in the pool.

k (required up)

Service requires at least k up out of n.

Interpretation

Common patterns:

Single component: k=1, n=1
Active-active pair: k=1, n=2
Quorum 2-of-3: k=2, n=3

Incident-based

Compute downtime from incidents. Enter rows or paste a quick list.

Default impact (%)

Used when a row does not include impact.

Paste incident list (optional)

Each line: minutes or HH:MM, optionally comma impact%. Rows below are ignored if you paste here.

Incident rows

Duration (minutes)	Impact (%)

Formula used

Service-level availability

Availability = (T − D) / T

Where T is the denominator window, and D is effective downtime.

Impact weighting

D_effective = D_minutes × (impact% / 100)

Helps represent partial outages or reduced capacity.

Reliability model

A_node = MTBF / (MTBF + MTTD + MTTR)

System availability uses a k-of-n pool:

A_sys = Σ C(n,i) A^i (1−A)^(n−i), i=k..n

SLA error budget

Allowed downtime = (1 − SLA) × T

Remaining budget = allowed downtime − effective downtime.

Approximate “nines”

nines ≈ −log10(1 − availability)

Useful for quick comparisons across systems.

How to use this calculator

1) Select your method

Use service downtime if you have monitoring data. Use reliability if you have MTBF and MTTR. Use incidents for postmortem logs.

2) Set the measurement period

Pick a standard window or define a custom duration. If your SLA excludes maintenance, enable “Exclude planned maintenance”.

3) Add downtime inputs

Enter downtime minutes, or incident rows, or MTBF/MTTR components. Use impact to model partial outages.

4) Compare against your SLA

The calculator reports effective downtime, allowed downtime, and remaining error budget for the chosen window.

5) Export for reporting

After calculating, use the CSV or PDF buttons for audits, reliability reviews, and stakeholder updates.

Example data table

A compact reference table for common SLA levels and their typical allowed downtime.

SLA (%)	Allowed downtime per year	Allowed downtime per 30-day month	Allowed downtime per week
99.000	3 days, 15 hours, 36 minutes	7 hours, 12 minutes	1 hour, 41 minutes
99.500	1 day, 19 hours, 48 minutes	3 hours, 36 minutes	50 minutes
99.900	8 hours, 46 minutes	43 minutes	10 minutes
99.950	4 hours, 23 minutes	22 minutes	5 minutes
99.990	53 minutes	4 minutes	1 minute
99.999	5 minutes	0 minutes	0 minutes

Sample scenario

Window	Unplanned downtime	Planned maintenance	Maintenance excluded	Computed availability
30-day month	42 minutes (impact 100%)	60 minutes	Yes	99.90278%

In this sample, denominator time is 43,200 − 60 = 43,140 minutes. Availability = (43,140 − 42) / 43,140.

Availability, Uptime, and the Denominator Window

Availability is calculated from a time window T and effective downtime D. For a 30‑day month, T is typically 43,200 minutes. If your SLA excludes planned maintenance, subtract maintenance from T so you do not penalize approved change windows. This calculator shows both the total window and the denominator window, which helps align reliability reports with contract language and on‑call expectations.

Impact Weighting for Partial Outages

Not every incident is a full outage. A degraded API, regional brownout, or capacity drop can be modeled with an impact percentage. Effective downtime becomes D_effective = downtime × impact. For example, 40 minutes at 60% impact counts as 24 effective minutes. Using impact keeps the availability number honest while still reflecting customer experience when some traffic succeeds or service remains partially usable.

Reliability Inputs: MTBF, MTTD, MTTR

When you have engineering reliability data, use the MTBF/MTTD/MTTR method. Node availability is A = MTBF / (MTBF + MTTD + MTTR). Lowering detection time often improves availability as much as speeding repairs, because MTTD adds directly to restoration time. Track MTTR for both automated recovery and human intervention, then compare modeled results to observed downtime to validate assumptions.

Redundancy and k‑of‑n Service Requirements

Many platforms survive single failures through redundancy. The k‑of‑n model estimates system availability from identical component availability. Active‑active pairs commonly map to k=1, n=2, while quorum systems may be k=2, n=3. If A_node is 99.0%, a k=1, n=2 pool yields about 99.99% availability, because the service is down only when both components fail simultaneously.

SLA Targets and Error Budgets for Engineering Decisions

SLA targets convert quickly into an “error budget” that guides release velocity and risk. Allowed downtime is (1 − SLA) × T, and remaining budget is allowed minus effective downtime. If a team consumes budget early in a month, freeze risky changes, reduce blast radius, and improve monitoring. If budget stays healthy, you can ship faster while still meeting contractual expectations. Use the downtime equivalents table to translate a percentage into minutes per week, month, and year for stakeholder communication and planning.

FAQs

1) What is the difference between uptime and availability?

Uptime is time the system is running. Availability is the fraction of the agreed window where users can successfully use the service, including partial outages and SLA exclusions.

2) Should planned maintenance count as downtime?

Only if your contract or SLO includes it. If maintenance is excluded, subtract planned minutes from the denominator so change windows do not reduce the reported availability.

3) How do I choose between downtime and reliability methods?

Use downtime when you have measured outage minutes from monitoring or incident reviews. Use reliability when you want a forward-looking estimate from MTBF, detection time, and repair time, especially during design or capacity planning.

4) How does the impact percentage affect results?

Impact scales downtime to represent degraded service. A 50% impact means each minute counts as half a minute of effective downtime, which changes both availability and remaining error budget.

5) What does “approximate nines” mean?

It converts availability into a shorthand using −log10(1−A). It is approximate because it ignores traffic shape, correlated failures, and distribution tails, but it helps compare systems with different uptime percentages quickly.

6) Can I model multi-region or clustered services?

Yes. Use k-of-n with n regions or nodes and set k to the minimum that must be healthy to serve traffic. For asymmetric regions, run scenarios per region and compare results using the export outputs.