Reliability & Availability

Preview — 3 of 10 questions

What is the correct hierarchy: SLI → SLO → SLA?

javascript

SLI (Service Level Indicator) — the metric itself:
  "99.7% of requests in the last 30 days returned 2xx in under 200ms"
  Raw measurement from your monitoring system

SLO (Service Level Objective) — your internal target:
  "We aim for ≥ 99.9% of requests to succeed under 200ms"
  Internal goal the team commits to; alerts fire when at risk

SLA (Service Level Agreement) — the customer contract:
  "We guarantee 99.5% uptime per month. If we miss it, you get credits."
  External commitment with financial consequences for breach

Why SLO > SLA?
  SLO (99.9%) > SLA (99.5%) — you aim higher internally so breaching
  the SLO triggers an alert before you breach the SLA and owe credits.
  The gap between SLO and SLA is your "error budget buffer."

ASLI (indicator) is the actual metric you measure; SLO (objective) is your internal target for that metric; SLA (agreement) is the external contract with customers, usually with financial penalties for breach.

BSLA defines the metric measured; SLO sets the internal target; SLI is the legal contract.

CAll three are the same concept with different names used by different cloud providers.

DSLO is the metric measured; SLA is the internal goal; SLI is the customer contract.

A service has a 99.99% availability SLA. How much downtime per year is permitted?

A~52 minutes per year

B~8.7 hours per year

C~5 minutes per year

D~26 seconds per year

What is the key operational difference between active-passive and active-active failover?

javascript

Active-Passive:
  Primary ──▶ handles all traffic
  Standby    ← ready but idle (hot standby) or cold (needs boot)

  Primary fails:
  → Health check detects failure
  → DNS or VIP (Virtual IP) switches to Standby
  → Failover time: seconds (hot) to minutes (cold)
  → Standby was wasted capacity while Primary was healthy

Active-Active:
  Server A ──▶ handles 50% of traffic
  Server B ──▶ handles 50% of traffic

  Server A fails:
  → Load balancer detects failure
  → Server B handles 100% of traffic
  → No downtime (already serving)
  → Both servers utilized during normal operation

AIn active-passive both servers process requests simultaneously; in active-active only one server is on at a time.

BActive-active is only available for stateless services; active-passive works for all services.

CActive-passive requires more hardware than active-active.

DIn active-passive the standby server is idle until the primary fails; in active-active all servers handle production traffic simultaneously, providing both redundancy and load distribution.

Answer all 10 questions (7 more), see explanations for every answer, and track your score.