MTBF, MTTR and Availability Calculator

What this calculator measures

Every maintenance and reliability program rests on three numbers: how long equipment runs between failures, how quickly it is restored when it does fail, and what fraction of the time it is actually available to work. Mean time between failures captures reliability, mean time to repair captures maintainability, and availability combines the two into the single figure a plant is ultimately judged on. This calculator turns your operating and repair data into all three, and it goes further than most by separating the three distinct kinds of availability, because the difference between them is usually where the real uptime problem is hiding.

The distinction matters because a machine can look excellent on paper and still deliver poor uptime in practice. Inherent availability, computed from MTBF and MTTR alone, is the best the equipment can do once a failure happens and only corrective repair is counted.

But real downtime includes planned maintenance and, more importantly, the waiting: for a technician, for a spare part, for a purchase order to clear, for the next scheduled window. Achieved availability folds in the planned maintenance, and operational availability folds in all the delay, and it is operational availability that an operator actually experiences.

Reporting only the inherent figure is how a maintenance dashboard can show ninety-five percent while the line is really down far more often.

This tool computes MTBF, MTTR, the failure rate, and all three availabilities from either raw maintenance logs or a datasheet MTBF and MTTR, then translates the result into an annual downtime budget and the familiar uptime nines. Everything runs in your browser, nothing you enter is stored, and each result comes with a chart and export so it travels from the plant floor to a review meeting intact.

How to use this calculator, step by step

Begin by choosing an input mode. If you have maintenance records, keep the logs mode and enter the total operating time over the period you are analysing, the number of failures in that period, and the total corrective repair time spent on them. From just those three the calculator gives MTBF, MTTR, the failure rate, and inherent availability. If instead you have an MTBF and MTTR from a supplier datasheet or a specification, switch to direct mode and type them in to get the same metrics immediately.

To go beyond the inherent figure, add the optional downtime fields. Enter the planned preventive-maintenance downtime over the period to get achieved availability, and enter the logistics and administrative delay, the time spent waiting for parts, crew, or authorisation, to get operational availability. As you add each one the governing availability shown at the top drops to the most realistic value your data supports, and the chart adds a bar so you can see the size of each loss. Every field recomputes live, so you can immediately see how shaving repair time or cutting parts-waiting would move the number.

The result panel leads with the governing availability, then lists MTBF, MTTR, the failure rate, and the inherent, achieved, and operational availabilities side by side, followed by the annual downtime and the uptime nines. The chart compares the three availabilities so the gap between design and reality is obvious at a glance, and you can download the analysis as a PDF or CSV or share it. The tool opens with a worked example already filled in so you see a complete, correct result before changing anything.

Reading MTBF, MTTR and the failure rate

MTBF and MTTR are simple averages, and understanding exactly what they average keeps you from misreading them. MTBF is total operating time divided by the number of failures, so it is an average of the good stretches between breakdowns; it says how often the equipment fails but nothing about the consequences when it does.

MTTR is total corrective repair time divided by the number of repairs, an average of how long the fixes take; it measures your maintainability, the combined speed of diagnosis, parts, and labour.

The two are independent levers: you raise availability either by failing less often, which lifts MTBF, or by recovering faster, which lowers MTTR, and knowing which lever is cheaper to pull is often the most valuable thing the analysis reveals.

The failure rate is the same information as MTBF seen from the other side. Under the constant-hazard assumption, the standard baseline for reliability work, the failure rate is exactly one divided by MTBF, so an MTBF of a hundred hours is a failure rate of 0.01 failures per hour. The failure rate is the more convenient form when you combine components, because independent failure rates add, and it is the direct input to the exponential reliability function that predicts the chance of survival to a given age. This calculator reports both so you can hand the failure rate to a system-reliability model or quote the MTBF to a manager, whichever the audience expects.

One caution keeps these numbers honest: MTBF describes repairable equipment, where the item runs, fails, is repaired, and runs again. For a part that is simply replaced when it fails rather than repaired, the right measure is mean time to failure, and while the two are numerically close when repair time is short, they answer different questions. If your interest is the life of a component up to its single terminating failure, and especially if failures cluster early or late rather than at a constant rate, Weibull life analysis is the appropriate tool rather than a constant-rate MTBF.

The three availabilities, and why the gap matters

Availability is deceptively simple as a formula and surprisingly subtle in practice, because the answer depends entirely on which downtime you agree to count. Inherent availability counts only the corrective repair that is unavoidable once a failure occurs, and it equals MTBF over MTBF plus MTTR. It is the cleanest measure of the equipment itself, the number a manufacturer can stand behind, but it is also the most optimistic, because it silently assumes that the instant a machine fails a technician is present with the right part and repair begins immediately.

Achieved availability relaxes that first idealisation by adding planned preventive maintenance to the downtime, giving the view a maintenance department holds when it counts both breakdowns and scheduled service. Operational availability relaxes the rest: it adds every logistics, supply, and administrative delay, the hours and sometimes days a failed machine sits waiting for a part to arrive, a contractor to be scheduled, or a work order to be approved.

Operational availability equals total up time over total time and is what the operator on the floor actually lives with. In many real plants it is far below the inherent figure, and the whole value of separating the three is that the gaps tell you where to act: a large inherent-to-achieved gap points at excessive planned downtime, while a large achieved-to-operational gap points at a spares and logistics problem no amount of faster wrenching will fix.

This is why quoting a single availability number without saying which one is a common and expensive mistake. A supplier may promise inherent availability while the customer hears a guarantee about real uptime, and contracts should specify which measure applies. The calculator makes the distinction concrete by showing all three from the same data, so you can see not just how available your asset is but which category of downtime is costing you the most.

Five worked examples you can follow

Example 1: the basic three metrics

The calculator opens with a thousand operating hours, ten failures, and fifty hours of repair. That gives an MTBF of a hundred hours, an MTTR of five hours, a failure rate of 0.01 per hour, and an inherent availability of about 95.24 percent. This is the irreducible core of the analysis, the numbers you can get from three fields, and it is worth changing the failure count or the repair time to feel how each one moves availability before adding anything more.

Example 2: adding planned maintenance

Now add thirty hours of planned preventive maintenance over the same period. The inherent availability is unchanged because the equipment fails no more often, but the achieved availability drops to about 92.59 percent because the machine is now also down for scheduled service. The gap between inherent and achieved is the price of your maintenance plan, and seeing it explicitly helps judge whether the plan is buying enough reliability to justify the downtime it consumes.

Example 3: adding logistics delay

Add seventy hours of logistics and administrative delay, the waiting for parts and crew. Operational availability now falls to about 86.96 percent, well below both other figures, and it becomes the governing number because it is the most realistic. The large achieved-to-operational gap is a clear signal that the constraint here is not repair skill but supply and scheduling, and that stocking a critical spare or streamlining approvals would recover more uptime than any change on the tools.

Example 4: the downtime budget in nines

Read the annual downtime and nines the calculator reports for the governing availability. An operational availability near 87 percent corresponds to well over a thousand hours of downtime a year, a figure that lands far harder than the percentage alone. Push the inputs toward three nines, 99.9 percent, and the annual downtime falls toward eight or nine hours; toward five nines and it is only minutes. Watching the downtime budget shrink as availability climbs is the most persuasive way to size a reliability target.

Example 5: working from a datasheet

Switch to direct mode and enter an MTBF of a hundred hours and an MTTR of five hours from a specification. You immediately get the inherent availability and failure rate without any log data, which is the common situation when evaluating a machine you have not yet run. Add a mean downtime figure that reflects your own expected parts and logistics delay, and the tool returns an operational availability that adjusts the vendor number to your reality, often a sobering correction to a glossy datasheet claim.

Three expert tips for a trustworthy number

The mathematics behind the results

The formulas are short, and seeing them makes the results easy to trust. MTBF is total operating time divided by the number of failures, and MTTR is total corrective repair time divided by the number of repairs, so both are ordinary averages over the period you supply. The failure rate is one divided by MTBF, the reciprocal that expresses the same reliability as failures per hour rather than hours per failure. These three follow directly from counting, with no distributional assumption beyond treating the period as representative.

The availabilities are all of the form up time divided by up time plus downtime, differing only in which downtime is included. Inherent availability uses only corrective repair, and because MTBF and MTTR are the per-failure averages, MTBF over MTBF plus MTTR is algebraically identical to operating time over operating time plus repair time.

Achieved availability adds the planned-maintenance downtime to the denominator, and operational availability adds the logistics and administrative delay as well, so the three form a nested sequence that can only decrease as more downtime is admitted. The annual downtime is simply one minus the governing availability multiplied by the hours in a year, and the nines are the number of leading nines in the availability percentage, computed as the negative base-ten logarithm of the unavailability.

None of this is heavy mathematics, which is precisely the point: the value of the tool is not in complex computation but in applying the right definition consistently and showing the three availabilities together.

Two modelling assumptions are worth stating. The calculator treats the failure rate as constant over the period, which is the standard baseline and is exact for the exponential model but only approximate when equipment is in infant mortality or wear-out; for those regimes the Weibull analysis is the correct complement. And it treats the period you supply as representative, so a window that happens to contain an unusual cluster of failures or an unusually smooth run will bias the averages. Using a long enough and typical enough period is what keeps the numbers meaningful.

Where these metrics are used

MTBF, MTTR, and availability are the shared vocabulary of maintenance and reliability across every asset-intensive industry. In manufacturing they drive equipment effectiveness and the case for preventive maintenance; in data centres and IT they underpin the service-level agreements written in nines; in transport, energy, and heavy industry they govern how fleets and plants are staffed, spared, and scheduled. Reliability-centered maintenance and total productive maintenance both use them as primary indicators, and asset-management standards such as ISO 55000 expect them to be tracked and improved. Wherever uptime has a cost, these three numbers are how that cost is measured and managed.

The metrics also connect directly to the rest of the reliability toolkit, which is why they are the flagship of this silo. The failure rate this calculator reports is the input to the exponential reliability function that predicts survival to a given age, and to the system reliability model that combines components in series and parallel.

MTBF and target availability feed the preventive-maintenance calculator that finds the cost-optimal service interval, and the failure modes that drive MTTR up are exactly what an FMEA ranks by risk. Availability sits at the centre of that web: it is the outcome the other tools are ultimately trying to improve, and reading it correctly, with the three types separated, is the starting point for using any of them well.

Return to the Maintenance and Reliability hub for the full set.

It is worth remembering that availability is only ever half of a business case; the other half is the cost of the downtime it measures. An hour of downtime on a bottleneck line that gates a whole plant is worth far more than an hour on a machine with spare capacity behind it, and the same availability figure can therefore justify very different maintenance spend depending on where the asset sits.

The practical habit is to pair the annual downtime this calculator reports with a cost per hour of downtime, turning a percentage into a currency figure that competes on equal terms with the cost of the reliability improvement being considered.

A ninety-five percent availability that costs little to lose may not be worth improving, while the same figure on a critical asset may justify a large investment in spares or redundancy, and only the monetised downtime makes that comparison honest.

Turning maintenance records into good inputs

The quality of the answer depends entirely on the quality of the inputs, and a little discipline in preparing them pays off. Operating time should be the time the equipment was actually running or ready to run over the analysis window, not simply the calendar span, because idle time that is not a failure should not inflate the denominator. The failure count should include only genuine failures that stopped the function, not minor adjustments or planned stops, since miscounting failures distorts MTBF directly. And the repair time should be the actual hands-on corrective time, kept separate from the waiting time, so that MTTR measures maintainability rather than logistics.

That separation is the crux of getting the three availabilities right. The single most common data problem is lumping all downtime together, which collapses the distinction between a slow repair and a long wait for parts and hides the very insight the analysis exists to provide. Record corrective repair, planned maintenance, and logistics or administrative delay as three separate buckets, and each availability falls out cleanly. If your records only have total downtime, you can still get inherent availability by using repair time alone, but you lose the ability to see whether your uptime problem is a maintenance problem or a supply problem.

Finally, choose the analysis window with care. It should be long enough to contain a reasonable number of failures, because an MTBF built on one or two failures is very uncertain, and it should be typical rather than a period distorted by a commissioning shakeout or a one-off disaster. When equipment is new or newly overhauled it is often in a changing-failure-rate regime where a single MTBF is misleading, and that is the signal to bring in Weibull analysis alongside this tool. Good inputs, consistently scoped, are what turn these simple formulas into numbers you can act on.

When MTBF and availability are not enough

These metrics are powerful, but they rest on assumptions that do not always hold, and knowing their edges keeps you from over-trusting a clean number. The biggest assumption is the constant failure rate. MTBF and the exponential model treat failures as equally likely at any age, which is a good description of the random-failure middle of an asset’s life but wrong at the ends: new equipment often suffers early-life failures that fall off with time, and old equipment suffers wear-out failures that climb. In those regimes a single MTBF averages over a changing reality, and Weibull analysis, which estimates how the failure rate varies with age, is the honest tool.

A second limit is that availability as computed here is a long-run average and says nothing about the pattern of downtime. An asset that is down for one long outage a year and one that is down for many short ones can share the same availability while having very different operational consequences, and neither MTBF nor availability distinguishes them.

Where the timing and clustering of failures matter, or where a single long outage is far worse than its duration suggests, availability needs to be read alongside the distribution of downtime, not on its own. Similarly, these metrics assume the failures are independent; a common cause that takes out several machines at once is not captured by combining their individual rates.

None of this diminishes the value of MTBF, MTTR, and availability for the routine job of measuring and improving uptime; it simply marks the point where a life-data or system model earns its extra effort, and recognising your situation before quoting a number is what keeps the number honest.

Common mistakes to avoid

A handful of errors recur and quietly distort the numbers. Watch for them.

• Quoting availability without saying which type. Inherent, achieved, and operational can differ by many points; a bare percentage invites the supplier-versus-customer mismatch this tool exists to prevent.
• Lumping all downtime together. Counting repair, planned maintenance, and parts-waiting as one number hides whether your problem is maintenance or logistics, the single most useful thing the split reveals.
• Mismatched periods. Operating time, failures, and downtime must all cover the same window; a year of failures against a month of running inflates the failure rate twelvefold.
• Counting non-failures as failures. Minor adjustments and planned stops are not failures; including them deflates MTBF and misstates reliability.
• Using MTBF on non-repairable parts. For items that are replaced, not repaired, the measure is MTTF, and if failures are not random, Weibull analysis, not a single MTBF.
• Trusting an MTBF built on one or two failures. A handful of failures gives a very uncertain average; widen the window until the count is meaningful.

Raising MTBF versus reducing MTTR

Because availability depends on both how often equipment fails and how fast it recovers, there are always two routes to more uptime, and choosing between them is one of the most practical decisions the analysis informs.

Raising MTBF means the asset fails less often, which you achieve through better design, higher-quality components, condition monitoring, or preventive maintenance that heads off failures before they happen.

Reducing MTTR means each failure costs less downtime, which you achieve through faster diagnosis, better spare-part availability, clearer procedures, and training. Both lift availability, but they cost very different amounts in different situations, and the numbers tell you which is cheaper.

A useful way to see the trade-off is to notice that availability is symmetric in the ratio of MTTR to MTBF. Halving MTTR and doubling MTBF move inherent availability by similar amounts when they change that ratio equally, so the question is purely which change is easier to buy.

For equipment that fails rarely but takes days to repair because a part must be fabricated, the leverage is all in MTTR and the answer is to stock the part. For equipment that is quick to fix but fails constantly, the leverage is in MTBF and the answer is to attack the root cause of the failures.

The calculator makes this concrete: change MTBF and MTTR in turn and watch which one moves availability more for a realistic amount of effort, and let that guide where the maintenance budget goes.

Setting and defending an availability target

Availability targets are easy to assert and hard to justify, and expressing them as an annual downtime budget is the discipline that keeps them honest.

Saying a line must hit ninety-nine percent sounds precise until you translate it into the roughly eighty-eight hours of downtime a year it permits, at which point the conversation becomes concrete: is eighty-eight hours acceptable, and what would it cost to halve it? Each additional nine is an order-of-magnitude reduction in downtime and usually a steep increase in cost, because it demands not just reliable equipment but fast repair and, above all, the elimination of logistics delay.

The jump from three nines to four is often far more expensive than the jump from two to three.

The three-availability breakdown is what makes a target defensible rather than aspirational. A target set on inherent availability is a statement about the equipment alone and can be met by buying a better machine; a target set on operational availability is a statement about the whole maintenance system, spares, staffing, and processes included, and usually cannot be met by equipment choice alone.

Confusing the two is how organisations commit to uptime they have no logistics plan to deliver. Use the calculator to work backwards from the downtime budget the business can tolerate to the availability it implies, then decide which of the three availabilities you are promising and check that the supporting spares and staffing make it reachable.

A target chosen this way comes with its own justification and its own action list.

A short history of these metrics

MTBF, MTTR, and availability grew out of the same mid-twentieth-century push that produced much of modern reliability engineering, driven first by military and aerospace demands where failures were expensive and often dangerous.

The United States military formalised reliability prediction in standards such as MIL-HDBK-217, which tabulated component failure rates so that a system rate, and hence an MTBF, could be estimated before anything was built, and the vocabulary of failure rate and mean time between failures spread from there into electronics, telecommunications, and eventually all of industry.

The availability formula, balancing reliability against maintainability, came with the recognition that keeping systems running was as much about fast repair and good logistics as about parts that did not fail.

The metrics reached their widest audience with the growth of computing and the internet, where the uptime nines became a marketing and contractual currency and five nines entered common speech as the gold standard for critical systems.

In parallel, the maintenance world adopted them as the backbone of reliability-centered maintenance and total productive maintenance, and asset-management standards drew them into formal governance.

What began as a way to predict whether a missile would work has become the everyday language of any organisation that depends on equipment staying up, which is why a firm grasp of these three numbers, and of the difference between the three availabilities in particular, remains one of the most portable skills in operations.

Input format and quick reference

Pick logs or direct mode. In logs mode enter operating time, number of failures, and corrective repair time, plus optional planned-maintenance and delay time. In direct mode enter MTBF and MTTR, plus an optional mean downtime. All times use the same unit (hours by default). The reference below explains each output.

Frequently asked questions

Sources, disclaimer and editorial transparency

This calculator applies the standard reliability definitions of MTBF (operating time per failure), MTTR (repair time per repair), failure rate (the reciprocal of MTBF), and the three availability measures: inherent (MTBF/(MTBF+MTTR)), achieved (adding planned maintenance), and operational (adding logistics and administrative delay), consistent with reliability engineering practice and asset-management standards such as ISO 55000. This calculator and guide are created and reviewed by the OpsCalculators team; see our Editorial Policy for how each tool is researched, built, and tested.

Results are accurate estimates for planning and education, not certified reliability or safety engineering advice, and they assume a constant failure rate over a representative period and independent failures. For equipment in infant-mortality or wear-out, use Weibull analysis; where the pattern of downtime matters, read availability alongside the downtime distribution. See our full Disclaimer. OpsCalculators.com is operated by MAFHH INTERNATIONAL LTD. Your data is processed in your browser and never stored; see our Privacy Policy.