SLA vs Support & Services
There are two layers to what a vendor promises you. The first is the contractual SLA: a number with a remedy. The second is everything else they sell to help you: premium support, designated humans, architecture reviews, professional services. Both are valuable. Only one is a guarantee. People mix them up all the time. Here is the line.
The SLA contractual
What the provider legally owes you when their uptime slips.
- Uptime commitment (e.g. 99.9% / month)
- Service credits with a defined % and cap
- Claim window (usually 30–60 days)
- A measurement method and a remedy if you file in time
Support & Services target / entitlement
What you can buy to get more help. Useful, often paid, almost never credit-backed.
- Support response times (Sev1 in 15 minutes, etc.)
- 24x7 coverage and faster escalation
- Designated humans: TAM, CSM, named architect, support team
- Architecture / Well-Architected reviews
- Professional services (sold via SOW)
- Customer success programs, training, enablement
Why this matters
If a provider's salesperson tells you "we have a 15-minute response SLA," that number is almost always a target sitting inside a support plan, not a credit-backed promise. A missed 15-minute response usually has no remedy beyond "we'll try harder," or at most "we'll re-perform the support." The genuinely enforceable promise — the one tied to money — is the uptime SLA above the support tier.
The reverse is also true. The premium support and services layer might be worth far more than the uptime credits ever will be — a Technical Account Manager who knows your architecture, a quarterly Well-Architected review, a 24x7 designated team for Sev1. Especially when the outage is on your side, not the provider's.
How to read a vendor doc in 30 seconds
| You see this wording | Treat it as |
|---|---|
guarantees, shall, with a Service Credit schedule + cap + claim window | SLA (credit-backed) |
target, objective, SLO, estimated, commercially reasonable efforts, every reasonable effort | Target / SLO (no remedy) |
included with X tier, designated TAM, access to architecture review | Entitlement (you can call on it, no measured outcome) |
A Statement of Work (SOW) with deliverables, milestones, acceptance criteria | Contractual deliverable (the one place "soft" services become measurable) |
What each role actually cares about
"What am I legally owed, and what does the insurance cost?"
- Uptime %, credit %, cap, exclusions
- Claim mechanics, window, automatic vs filed
- Premium-support cost as a % of spend, and what it buys
- Chronic-failure / termination rights — the real leverage
"Can this provider help us run it better, and respond fast when it breaks?"
- Sev1 response targets (treat as SLOs, not guarantees)
- Designated TAM/CSM/team — continuity and expertise
- Architecture / Well-Architected reviews
- 24x7 escalation, professional services for migrations
"Coverage, escalation, and accountable contacts."
- Response times per severity, business vs 24x7
- Designated / named contacts and escalation paths
- Authorized contacts, case-volume limits per tier
- Training, onboarding, enablement entitlements
How SLA.directory shows this
On every vendor page we keep the two layers clearly separated. The SLA report card at the top is the contractual layer: uptime, credits, claim window, premium SLA tiers. Below it, when the vendor publishes one, a Support & Services section lists the entitlements layer with each promise tagged as a target, an entitlement, or — rarely — a real SLA. We never label a support response time as an SLA unless the vendor's own document attaches a credit to it. If you see "Target / SLO" next to a number, that's because the vendor's wording said so.
The data is open: see the dataset, the methodology, or suggest a correction.