Lesson 3 of 4 · Free to read
Session guide
Session guide — Performance, SLOs & Data-store Signals
Visual companion to the Lesson. Use in study groups or when you want the diagram first.
Watch
Course video
Course video
Flow
60-second summary
One availability SLO on “the api service” can be green while users wait forever for async fulfillment. MNCs define SLIs per user journey and write error-budget policies before the incident.
MNC 2026: k6 (or similar) for load, journey-based SLIs/SLOs, multi-window burn-rate alerts, and first-class Postgres/Redis telemetry are standard SRE practice.
Chunks
- Journeys not services: Sync
/ordersacceptance vs async “order in S3” freshness are different promises. - SLI shapes: Availability, latency, freshness, throughput — all good/valid ratios.
- Policy beats vibes: 25%/50%/100% budget actions agreed in advance (freeze features, page next tier, etc.).
- Data stores: Connection pools, slow queries, Redis hit ratio — silent killers of freshness SLOs.
Session guide — Performance, SLOs & Data-store Signals.