Enrolment
This course is currently free to enrol in.
Public course
Enterprise AI — Course 15: Capacity, Performance and Reliability Design
Practical service design using demand, latency, capacity and recovery evidence to compare feasible AI service options.
AIcapacityperformancereliabilityqueuesservice objectivesrecovery
By Telari Labs
Enterprise AICourse 15 of 25
Learning on Telari
Created by Telari Labs, delivered through Telari
The course creator controls the material and assessment standards. Telari provides structured lessons, practice, feedback against those standards and a record of your progress. See how Telari works.
AI access
The first managed lesson is included on Free. Continue with Plus or your own supported API key; premium assessments and certificates may require Plus.
Audio
No reviewed lesson audio is currently advertised for this course.
Lesson list
- Demand and service unitsWorkload estimates that distinguish task demand, model calls, tokens and concentrated arrivals.
- Latency and throughputUser-relevant deadlines, complete traces and performance distributions.
- Concurrency, queues and limitsBounded capacity estimates, backlog behaviour and overload decisions.
- Capacity and operating arrangementsBottleneck-specific capacity choices and the responsibilities needed to sustain them.
- Performance interventionsImprovements that preserve the task's freshness, quality and completion requirements.
- Dependency failures and availabilityEnd-to-end service availability and the limits of nominal redundancy.
- Recovery objectives and degraded serviceRestoration targets, recoverable state and honest restricted operation.
- Designing load and recovery experimentsRepresentative experiments with useful observations and bounded authority.
- Capacity triggers and uncertaintyReassessment tied to changed assumptions and useful decisions.
- Choosing a feasible service designEvidence-based recommendations within quality, resilience and operating constraints.
- Final service design assessmentIntegrated written decisions about demand, performance, resilience and feasible service operation.