Real World SLOs and SLIs: A Deep Dive
summary
Defines SLOs from the outside in by mapping system boundaries to concrete capabilities, then measuring SLIs (eg. 99.9% uptime target for platform ingest and compound SLOs across insert/ query behavior). Uses New Relic examples, browser monitoring, Mesos/Marathon task states and synthetic scripts to measure startup and end‑to‑end workflows, sets SLIs on the 95th–99th percentile rather than averages, treats hard‑sharded shards as separate SLOs (one SLO per failure domain), and warns not to page on raw SLI violations while iterating SLOs over time (use synthetic scripts).
AI-written summary · about
chapters (15)
- 0:00Intro
- 2:30System Boundaries
- 6:23Availability
- 10:18RDDB
- 13:31Hard Sharding
- 16:45Network Tier
- 17:42Dependencies
- 20:18Container Scheduler Cluster
- 20:47How to Define a SLO
- 23:59UI Tier
- 25:21Browser Monitoring
- 25:55Sanity Check
- 26:59Customer Experience
- 28:36Summary
- 29:59Questions
related talks
32m9 Things You Should Do When Starting to Use SLOs
38mImplementing SRE practices: SLI/SLO deep dive
20m9 SLIs; Oh My!
44mHow to Extract SLOs: Sources and Tools of Midband SRE Gnomes
57mSLI and SLO for the Business: Monitoring Quality Across 200+ Products
18mSLICK: Driving SLO Culture
23mGoing from 30 to 30 Million SLOs
18mSLOs as an Organisational 'Check Engine' Light