Saying "the site should be reliable" is too vague to act on, so teams agree on a number. An SLI is what you measure, an SLO is your internal target for it, and an SLA is the promise to customers with a penalty attached. Use them when deciding how much downtime is acceptable, for example "99.9% of requests succeed over 30 days". People often mix up SLO and SLA; the SLA is usually looser so the team is warned before money is at stake.