On call does not have to suck
summary
Argues that owning code and redesigning alerts makes on‑call sustainable: adopt event‑level observability, feature flags and auto‑deploy to a small percentage, track DORA metrics (deploy frequency, lead time, MTTR) and aim for SLO/SLS‑aligned alerts. Gives concrete wins: a chaos habit — kill the oldest coffee node every week — fixes bootstrapping; high performers deploy ~973× more frequently and recover faster. Recommends volunteer rotations, manager backup shifts, and compensating rough nights with time off.
AI-written summary · about
chapters (17)
- 0:00Intro
- 0:22What was the conference like
- 0:42Who is Charity
- 1:57Toxic Experiences
- 2:24Badge of Honor
- 3:46Ownership
- 4:35Operations
- 5:16Writing Code
- 5:50Technical Managers
- 10:03Out of Hours Alerts
- 10:17Embrace Real Observability
- 11:10Automatic Deployments
- 11:56Technical Debt
- 12:58How well does your team perform
- 13:33Metrics
- 25:33On Call
- 29:31Pay People
related talks
31mHow to Build a Healthy On-call Culture
06mIgnite: Key Factors for Successful 24×7 On-Call Duty
26mBuild a data-driven on-call workflow for your team with atomic habits
22mRaise a Production System:Making on-call human-scale w/ Observability & Trust
12mHow not to burn out your monitoring team
28mOn-Call Experience at Uber
33mTackling Alert Fatigue with SLOs
45mIncident Management - Talk the Talk, Walk the Walk