One serving replica has a per-token latency 40 percent worse than its peers. Find out why.

Unlike a training job, a slow serving replica holds nobody else back, so it survives unnoticed while degrading a slice of users. The five causes and the metric that distinguishes each, why routing can create the symptom with no hardware fault, and the fix that limits damage while you investigate.

Updated Sep 2026 · Grounded in real AI infrastructure interview loops and written to a senior-engineer editorial bar, with every number worked and every diagram hand-built.

20 answers per topic instead of 10, plus saved progress and bookmarks · no cardor unlock all 283 remaining answers · ₹2,000 / $25

The concepts behind this question

Ranked by how closely each one overlaps this question's topic, so the first card is the thing to read if the answer above moved too fast.

Foundational

🚀 Inference & Serving

Prefill vs DecodeAn LLM request runs in two phases with opposite hardware profiles: prefill reads the whole prompt in one compute-bound pass and decides time to first token, decode emits one token per forward pass and is bound by memory bandwidth. Every serving decision, from batch size to which GPU to buy to whether to split the two phases across machines, follows from that split.

Foundational

🧮 Open Weights & Serving Engines

Serving Benchmarks That Do Not LieMost published serving numbers are not comparable to each other and not predictive of production, because they differ in the input distribution, the concurrency, whether the cache was warm, and which of several very different metrics is being reported. A benchmark that supports a decision has to fix all four, report a distribution rather than a mean, and be run against the traffic shape you actually serve. The single most useful discipline is to compute the bandwidth bound first, so you know what fraction of the possible you achieved.

Core

🩺 Fleet Reliability & ObservabilitySign in

SLOs for AI SystemsA service level objective is a promise with a number attached, and AI systems need their own because the classic ones do not fit: a training run has no requests, only progress, so its objective is goodput; an LLM endpoint streams, so its latency is two numbers (time to first token and time per token) rather than one; and both spend a budget that is set by hardware failure rates rather than by software bugs. This page defines the objectives that fleet and serving teams actually use, derives the thresholds from user needs and from the hardware, and works the error-budget arithmetic that decides when to stop shipping and start fixing.

Advanced

🚀 Inference & Serving🔒 Premium

Chunked PrefillA long prompt's prefill can occupy a GPU for hundreds of milliseconds, and every sequence mid-decode on that GPU waits for it. Chunked prefill splits the prompt into fixed token budgets and interleaves each chunk with a decode step, so decode latency stays flat at the cost of a slower first token for the long prompt. The chunk budget is a knob between TTFT and TPOT, and the interview question is how you would set it.

Explore the full curriculum →

UP NEXT ON YOUR JOURNEY

Next in this trackWrite the policy for when a GPU is replaced rather than returned to service. What are the triggers and what do they cost?Next in this trackHow much spare capacity does a 16,000-GPU fleet need, and what are you actually reserving it for?Next in this trackHow would you build a postmortem practice that people take seriously and that actually reduces incidents?

FEDITOR'S NOTE

Scored on knowing a slow replica affects a slice of traffic rather than the fleet, on distinguishing a hardware cause from a routing-induced one, and on shedding traffic from it as the immediate action.

DISCUSSION · 0

No comments yet — be the first to share your approach.