aiengineering.guideaiengineering.guide

Glossary

Glossary

Every term, defined plainly and precisely, linked to the lessons that use it.

B

backpressure
Signalling upstream to slow down when a queue is filling, instead of dropping work silently.

C

circuit breaker
A switch that stops calling a failing dependency for a cooldown, so failures do not pile up.

F

fallback
A cheaper or degraded path taken when the primary one fails or is over budget.

I

idempotency key
A client-supplied id that lets a retried request be de-duplicated to one effect.

R

rate limiting
Capping how many requests a caller may make in a window, to protect a shared resource.

T

tail latency
The slow end of the latency distribution (e.g. p99) — what the unluckiest requests see.
token bucket
A limiter that refills tokens at a fixed rate; a request spends one token or is rejected.

Enter to go · Esc to close