Watch the explanation. Practice the skill.
Cost, latency, and trade-offs
Measure response time and usage without hiding retries or unknowns.
Captions appear here during playback.
Cost, latency, and trade-offs
Measure response time and usage without hiding retries or unknowns.
Video playback is unavailable. Read the descriptive transcript below, go directly to practice, or open the video file.
Source-based teaching example · Fictional data · Reviewed 2026-10-03. Microsoft Mark · locally synthesized. On a small screen, use fullscreen for diagram details or read the transcript.
Read the descriptive transcript
0:00 · What you will learn
Measure response time and usage without hiding retries or unknowns.
What appears on screen
Measure response time and usage without hiding retries or unknowns.
0:07 · The idea
Record what work was performed and how long it took. Include repeated attempts. Use the declared rate and units, and do not treat missing usage as zero.
What appears on screen
Record what work was performed and how long it took. Include repeated attempts. Use the declared rate and units, and do not treat missing usage as zero.
0:20 · One useful term
Latency. The time taken for a response or task, measured from a defined starting point.
What appears on screen
The time taken for a response or task, measured from a defined starting point.
0:29 · A worked example
The exercise uses an invented rate card. A retry repeats work even if the first answer was unusable. Ten timing samples also cannot reliably describe very rare slow responses.
What appears on screen
The exercise uses an invented rate card. A retry repeats work even if the first answer was unusable. Ten timing samples also cannot reliably describe very rare slow responses.
0:44 · Your turn to think
Does a failed answer mean the first attempt used no resources? Pause and decide before continuing.
What appears on screen
Does a failed answer mean the first attempt used no resources? Pause the video if you need more time.
1:02 · Why this answer works
No. Record performed work and any unknown usage. Keep toy calculations separate from actual provider bills.
What appears on screen
No. Record performed work and any unknown usage. Keep toy calculations separate from actual provider bills.
1:13 · Check 1
Include retries in measured work.
What appears on screen
Include retries in measured work.
1:18 · Check 2
State the rates, units and sample size.
What appears on screen
State the rates, units and sample size.
1:24 · Check 3
Compare models with the same surrounding workflow.
What appears on screen
Compare models with the same surrounding workflow.
1:30 · Now try a different case
Open the lesson and try its different practice case. Use every check in the full task. Watching alone does not mark the lesson complete.
What appears on screen
Open the lesson Try the new case Check and revise The full task and its requirements are in the linked lesson.
Watching explains the example. Complete the lesson practice to check what you can do. Explore all visual lessons →