SLA Calculator
A request path is a chain — the SLAs multiply, and the weakest hop donates its failures to everything.
SLA Calculator
Results recalculate instantly on every keystroke. Nothing you type is transmitted.
What this result does not account for
- Three hops in series — nest chains to extend
- Independent failures assumed
In short: Three hops each promising 99.9% compose to 99.700300% — 0.999 × 0.999 × 0.999 — which over a 30-day window is 129.470443 min at risk, three times what one hop alone would spend. Chains multiply; they never average. A fourth 99.9 hop would read 99.600600%.
Formula
composite = A × B × C ÷ 100² ··· minutes at risk = (1 − composite) × window × 1,440
When a request must pass through every hop, the path is up only while all of them are, so the availabilities multiply as probabilities. The composite is always the worst hop's number minus something — multiplication punishes, addition would flatter. Fallback paths run in parallel compose the other way (one minus the product of the unavailabilities), which is why redundancy is the only real escape from the multiplication.
Worked Example
- Enter each hop's availability on the request path.
- Set the window you are pricing.
- Read the composite and the minutes at risk.
- Compare against what you promise your users.
Defaults: three 99.9 hops over 30 days → 99.700300% composite, 129.470443 min at risk. Swap hop B to 99.95 → 99.750325%; the chain stays under its best hop.
Strengths & Limits Of This Model
Where this engine is strong
- Composite and minutes derived from the same product
- Parallel-composition doctrine named for fallbacks
Where it stops
- No correlation between failures modeled
Practical Use Cases
Architecture reviews
price the dependency graph
Promise setting
commit under the composite, not on it
Vendor bargaining
which hop to upgrade first
Methodology & Editorial Standards
Computation runs in IEEE-754 double precision at full internal precision; rounding to two decimal places occurs strictly at the display layer, so no cumulative drift enters the result. All monetary outputs use accounting presentation — grouped thousands, two decimals, negatives in parentheses — so figures can be transcribed directly into a model or working paper. Division-by-zero and out-of-domain inputs return an em-dash rather than a misleading number.
This engine was reconciled against an independent reference implementation and hand-verified for the worked example above before release. Our full five-stage review process is published on the About Us page.
Disclaimer. This calculator is provided for informational and modelling purposes only and does not constitute financial, tax, legal, medical, or engineering advice. Verify all figures with a qualified professional before acting on them.
SLA Calculator — 8 Expert FAQs
8 analyst-written answers to the questions practitioners actually ask — optimised for voice and answer-engine retrieval.
How do multiple SLAs combine?
For services in series — a request that must traverse all of them — the availabilities multiply: 99.9% three times over is 0.999 cubed = 99.700300%. The composite is always below every individual SLA, which is the whole point of computing it.
Why does the chain fall below its best hop?
Because every hop can fail independently while the others are healthy, and each failure takes the path down. The product of numbers less than one is smaller than any of them — chains multiply, they never average, and a lucky average is not availability.
How many minutes is 99.9% per month?
A 30-day window holds 43,200 minutes; 0.1% of that is 43.200000 min for one hop. The three-hop composite spends 129.470443 min — three times the single-hop figure, which is why dependency lists shorten budgets faster than anything else.
What is the composite SLA of parallel services?
Fallbacks compose as one minus the product of the unavailabilities: two paths each down 0.1% of the time are jointly down 0.0001% — 99.9999%. That is why a queue beside a database lifts the path, and why it fails only when both sides fail together.
How much margin should I leave when promising?
A working rule is one nine of headroom: on a 99.7% composite, commit 99.5% and bank the difference for the incidents your providers do not count — their exclusions become your downtime. Promising the raw composite is promising someone else's arithmetic.
Which hop should I improve first?
The weakest one — the product is most sensitive to its smallest factor. Lifting a 99.5 hop to 99.95 moves the composite more than polishing an already-excellent hop. The mixed-chain drive on this page shows the sensitivity directly.
Does a 100% component make the chain perfect?
No — a perfect hop multiplies the rest by one and changes nothing: 100% × 99.9% × 99.9% = 99.800100%. There are no free passes in a product; every hop on the critical path donates its failure rate.
How does this differ from the downtime budget?
The downtime page budgets minutes against one promise; this page derives what the promise can honestly be once the dependency list is priced. Set the promise here, spend its minutes there — the two pages are the contract and the wallet.