SLA (Service Level Agreement)
Last updated
Definition
A service level agreement is the contractual attachment to a managed services agreement that defines what you support, the measurable targets you commit to – above all response time by ticket priority – and the remedy the client receives when you miss. It differs from an SLO: an SLO is an objective you measure and report, while an SLA is a promise with a penalty attached.
Why it matters to an MSP
The SLA is where your help desk stops being a best effort and becomes an enforceable obligation. The key distinction is response versus resolution. Response – a technician acknowledges the ticket and starts work within a set window, typically 15–30 minutes for a P1 outage and four to eight business hours for a P4 request – is under your control and belongs in the contract. Resolution depends on vendors, parts, and client cooperation, so treat it as a published target you report on, not a guarantee; commit to resolution times and you're in breach every time a carrier is slow.
Priority is yours to assign, not the client's, and an active security incident is a P1 under any sane matrix. Remedies are normally service credits of 5–20% of the monthly fee, applied automatically and written as the sole and exclusive remedy – the clause that keeps a missed response from becoming a damages claim. Measure attainment in the PSA, with the clock paused while a ticket waits on the client, and show the numbers at every QBR. Publish business-hours targets you can hit in your worst week, and sell true 24/7 as a premium tier only once you have the staff. The full priority matrix and contract language are in SLA design.
Related terms: SLO, Escalation, PSA, First Call Resolution