What does "high priority" mean? The secret to happy queues
summary
Argues that queues succeed when named as latency guarantees, not vague priorities, and shows an implementation: create queues like within 1 minute or within 10 minutes, enforce time limits (rule of thumb: job runtime ≈ one tenth rule of latency), add alerts when guarantees breach, autoscale on queue latency starting at a 50% threshold, and move long jobs to slower queues. After reorganizing, 60 queues collapse to about 9, alerts and autoscaling run, and latency problems largely disappear.
AI-written summary · about
chapters (5)
related talks
40mResilient Design 101: Queue Theory
30mThe Queue Continuum: Applied Queuing Theory
40mServiceability under high demand
38mA Practical Look at Performance Theory
41mGetting Out of Quicksand, With DevOps!
20mBuilding Reliable LLM Systems
56mDefending against CDD: Chaos-Driven Delivery
35mIs efficiency a good thing?