What should I check in the SLA before signing a SaaS contract?
The short answer
Before signing a SaaS contract, the SLA (service level agreement) deserves a careful read because the uptime percentage in the headline is rarely the whole picture. The practical terms that define what the commitment actually covers are the definitions: what counts as 'downtime,' what measurement window applies, which events are excluded from the calculation, what credit thresholds exist, and how support response time commitments are defined and measured. A 99.9% uptime commitment with broad maintenance exclusions and a monthly measurement window may provide materially less assurance than the headline suggests. Scan the SLA — often in an exhibit or addendum — before signing, not after an outage.
No account requiredFile deleted after analysisNot legal advice
What a SaaS SLA typically covers
An SLA in a B2B SaaS agreement typically covers at least two areas: system availability (the uptime commitment) and support response times (how quickly the vendor must respond to and resolve different categories of issues). Uptime is defined as a percentage — commonly 99.9% or 99.5% — measured over a window (monthly is most common; annual is less protective because failures early in the year can be averaged against clean months later). The SLA specifies how the vendor reports uptime, how the customer submits credit claims, and how credits are calculated and applied.
Support response time SLAs define severity tiers — typically critical (system unavailable), high (major functionality impaired), medium (functionality degraded), and low (cosmetic or minor issues) — and commit the vendor to responding within defined windows for each. Whether response time commitments are also tied to resolution time (not just initial response) is a practical distinction: a vendor that responds in 15 minutes but takes 72 hours to resolve has technically met a response SLA.
The definitions that change what the SLA actually means
The uptime percentage and the credit schedule are the two numbers most buyers focus on. The definitions that affect how those numbers play out are: what constitutes 'downtime' (full outage only, or does partial degradation count?), what events are excluded from the measurement (scheduled maintenance windows, third-party infrastructure failures, denial-of-service attacks, force majeure events), and whether the measurement is based on vendor-side monitoring, customer-reported outages, or a shared monitoring tool. Broad exclusion lists shrink the practical uptime commitment significantly; narrow definitions of 'downtime' may mean that a service operating at 50% capacity is not counted as an SLA failure.
What to look for in your agreement
- Uptime definition: does it cover only full outages, or does partial service degradation below a defined threshold count?
- Measurement window: monthly or annual — and who provides the uptime measurement data?
- Scheduled maintenance exclusions: how much scheduled downtime is permitted, when is it scheduled, and how much advance notice is required?
- Credit thresholds and amounts: what uptime percentage triggers credits, how large are the credits, and is there a maximum credit per period?
- Support response time SLAs: are response times defined per severity tier, and do resolution time commitments also exist?
Questions to ask before signing
- Ask the vendor to confirm the historical uptime for the specific service tier you are purchasing — not the SLA commitment, but the actual performance record.
- Ask the other party to clarify whether partial degradation (not a full outage) counts toward SLA calculations.
- Confirm the maximum scheduled maintenance window per month and the advance notice required.
- Consider having the SLA provisions reviewed alongside the limitation of liability and sole remedy clauses to understand the full picture of what you can recover if the SLA is missed.
Why scan instead of guess
The general rule tells you the baseline. Your agreement tells you what you’re actually being asked to sign — and the wording is what binds. Dang reads the document and flags the clauses worth reviewing, in plain English.
The deterministic engine scores and decides what’s risky. The AI only enriches the plain-English wording — AI extracts, code decides, never the other way around.
Your original file is deleted promptly after processing — we keep only the report you can read. No account needed for a one-time scan. Free preview first; full report $6.99, one-time.
Common questions
Is a monthly or annual SLA measurement window more protective for the buyer?
Monthly measurement is generally more protective. An annual window allows the vendor to average strong months against a month with significant downtime, potentially avoiding a credit obligation even for a sustained outage period. Monthly measurement means each month is evaluated independently, and a bad month triggers credits even if the rest of the year was clean.
What is a scheduled maintenance exclusion and why does it matter?
Most SLA uptime calculations exclude downtime attributable to scheduled maintenance. If the vendor reserves the right to take the system offline for multiple hours per week for maintenance, those hours do not count against the 99.9% commitment. How much maintenance is permitted, when it can be scheduled, and how much advance notice is required are all worth checking in the SLA before signing.
No account required · File deleted after analysis · Not legal advice. Dang reports contract findings in plain English — general information, not legal advice about your situation. For consequential decisions, consult a licensed attorney in your state.