Choose the two incorrect options regarding SLOs (service level objectives) and SLAs...

Option A is incorrect because legal liability arises from breaching the SLA (a contractual agreement), not the SLO (an internal engineering target). Option C is incorrect because SLAs are typically set at values that are easier to achieve than SLOs—SLAs are customer-facing commitments with consequences, while SLOs are more ambitious internal targets. Options B and D are correct: SRE teams use error budgets to intentionally cause controlled outages within SLO bounds for testing, and SLAs are indeed set by considering competitive stability, engineering feasibility, and customer requirements.

Step-by-step Derivation:
SLO vs SLA Definitions:

• SLO: Internal engineering target (e.g., 99.9% uptime)
• SLA: Contractual promise to customers with legal/financial consequences

Analyzing each option:

A) "Legal liability if SLO breached" — INCORRECT
→ Legal liability only applies to SLA breaches (contractual).
→ SLO breaches are internal and trigger error budget management, not legal action.

B) "Intentional outages within SLO bounds" — CORRECT
→ This describes error budgeting in SRE.
→ If SLO = 99.9% availability, the 0.1% downtime is the error budget.
→ SRE teams intentionally cause controlled outages (chaos testing) to stay within the budget.

C) "SLAs harder to achieve than SLOs" — INCORRECT
→ Relationship is reversed.
→ SLA is the customer commitment (easier to keep and defensible).
→ SLO is the internal engineering target (usually stricter/harder to achieve).
→ Example: SLA = 99% uptime, SLO = 99.9% uptime.

D) "SLAs consider competition, feasibility, and customer needs" — CORRECT
→ All three factors are standard inputs to SLA negotiation.
→ Competitive benchmarking, technical feasibility, and customer requirements balance the final SLA value.