Intermediate SRE / DevOps #latency #throughput #collocations

Latency & Throughput Collocations

5 exercises on the language engineers use to describe how fast and how much: latency, percentiles, throughput, round-trip time, and the hot path.

Key patterns
  • reduce / cut / shave latency; measure p95 / p99 tail latency
  • increase / boost / scale throughput, in requests per second
  • reduce round-trip time (RTT); collapse round trips
  • keep work off the hot path
  • latency = time per request; throughput = requests per time
0 / 5 completed
1 / 5
A platform engineer writes: "After adding the cache, we managed to ___ p99 latency from 800ms to 120ms."
Which verb is the standard collocation for bringing latency down?