Telari Labs

Student

Back to discover courses

Public course

Enterprise AI — Course 10: Reliable Actions and Durable Execution

Execution decisions grounded in evidence, covering retries, crashes, conflicting work and recovery while preserving useful completed actions.

executionretriesidempotencycheckpointsconcurrencyreconciliationcompensationhandoffs

By Telari Labs

Enterprise AICourse 10 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.

Enrolment

This course is currently free to enrol in.

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

  1. States and Execution EvidenceExecution records distinguish request progress, attempt failure and business outcomes and support justified state changes.
  2. Retries and IdempotencyRepeat-attempt decisions depend on intention identity, payload rules, deduplication scope, retention and explicit retry limits.
  3. Checkpoints and ResumptionDurable intent and verified effects support resumption within crash windows, local transaction boundaries and current authority.
  4. Concurrency and OrderingCompeting and delayed work require controls enforced where shared state changes.
  5. ReconciliationClaim-specific authority, correlation, query scope and explicit visibility guarantees support reconciliation of conflicting or missing records.
  6. Compensation and Irreversible EffectsRecovery for completed effects follows current authority and dependencies while accounting for uncertainty and remaining consequences.
  7. Handoffs and Bounded ContinuationResponsibility transfers through a stated contract; repeated delivery and continuation limits require explicit evidence and an accountable owner.
  8. Recovering a Partial WorkflowA worked recovery preserves verified effects, resolves the controlling uncertainty and changes the remaining actions when authoritative evidence changes.
  9. Revising Execution and Recovery DecisionsIndependent cases assess revised execution judgements and bounded recovery from supplied evidence.