IT Glossary
What is an SLA (Service Level Agreement)?
The service level agreement: a provider's promises turned into measurable numbers — availability, reaction times, remedies for falling short.
"We are always there for our clients" is marketing; "we respond to critical incidents within 30 minutes, monthly availability at least 99.9%, otherwise a 10% credit on that month's invoice" is an SLA — the agreement that converts promises into measurable commitments with consequences. The anatomy of a useful one, by the pieces that matter. The indicators: availability (uptime — the percentage of time the service works, with arithmetic worth feeling once: 99.9% monthly allows roughly 43 minutes of accepted downtime per month, 99.99% under five; each extra "nine" makes the service seriously more expensive, so it is bought against real need, not pride) and reaction times per severity level (critical/major/minor — with the key distinction weak contracts blur: response time means "we have started", resolution time means "it works again"; the first is easy to guarantee, the second is what counts). The definitions: what exactly qualifies as a "critical incident", how availability is measured (by whom, with which tool; exclusions for announced planned maintenance are normal — vaguely worded ones are holes), and over which window (24/7 or business hours only — an essential difference for a store that sells at night). The remedies: service credits for missed targets (percentages of the invoice — enough to sting at least symbolically; an SLA with no consequence is a brochure with a signature) plus, for repeated breaches, the right to terminate without penalties. Where you meet SLAs and where you should demand them: hosting and cloud (standard), website and application maintenance (the most commonly badly negotiated — ask for times per severity, not "we treat it with priority"), external IT support, and important SaaS products (their SLAs are published — read them before you depend, not after the incident). And the closing honesty from the other side of the table: an SLA does not guarantee the absence of incidents — it guarantees how they are answered and who bears the consequences; a vendor with a realistic SLA and a transparent compliance record beats one promising 100% behind foggy definitions — the first told you the truth before you had to discover it.
Let’s talk about your project
Message us on WhatsApp or send an email — you talk directly to a developer.
office@northdan.com · +40 752 070 247
Why it matters for your business
Verifiable promises, not impressions
Times and availability become measured, reported numbers — the conversation with the provider runs on data, not on contradictory perceptions.
Guaranteed priority when it burns
Defined severities with committed times mean your critical incident does not queue behind cosmetic requests — contractually, not out of kindness.
Consequences that align interests
Credits for missed targets and exit clauses for repeated breaches keep the provider invested in quality — not merely in invoicing.
Frequently asked questions
What availability level should I ask for — and what does each "nine" cost?
Base it on your real cost of downtime: for a brochure site, 99.9% (43 minutes a month) is honest and affordable; for a store or an application the business lives on, 99.95-99.99%, with redundant architecture and a price to match. Each extra nine multiplies infrastructure cost — buy it on the arithmetic of "what we lose per hour down", not on the reflex of "we want the maximum". And ask for the real track record: a 99.99% promise from a provider that never publishes its uptime is just a pretty number.
Where do SLA problems most often hide?
Top three: vague definitions (an undefined "critical incident" means everything is "minor" when convenient), response time sold as resolution time ("we respond in 15 minutes" — and fix it whenever), and elastic exclusions ("planned" maintenance announced an hour ahead, "external causes" covering anything). Plus measurement: who does it and with what — ideally you too, independently, with monitoring costing a few euros; your own numbers turn any dispute into a short conversation.
My maintenance provider has no SLA — "don't worry, we'll handle it". Is that a problem?
It is a lottery with a good history so far — until the incident that lands during a holiday, a rush, or a client bigger than you: without written commitments, your priority is a function of the day's goodwill. The reasonable, non-confrontational request: response times across two or three severities, a clear emergency channel, periodic reporting — one annex page in the contract. A good provider signs it easily (they already work that way); one who refuses any number has elegantly told you what "don't worry" is worth.
Let’s talk about your project
Message us on WhatsApp or send an email — you talk directly to a developer.
office@northdan.com · +40 752 070 247