Reliability & Delivery · Operational Resilience

Scheduler Agent Supervisor

Separate coordination, remote work and supervision. Durable step state and completion deadlines allow a supervisor to detect failed or timed-out work and arrange retry, recovery or compensation.

workflowrecoverysupervision
The problem
How can a distributed multi-step process detect stalled work and resume or compensate it after failures?
GateSift summary based on Microsoft Azure Architecture Center guidance. The wording, visualization, Azure mapping and analyzer context are original GateSift material.
Original GateSift visualization

How Scheduler Agent Supervisor works

A scheduler coordinates work, agents execute remote steps and a supervisor detects stalled or failed progress.

Durable task state
Schedule, execute, supervise
Completed
Recovered / compensated
1

Persist each step, owner, status and completion deadline before remote work begins.

2

Let agents execute idempotent work while the scheduler advances the workflow.

3

Use the supervisor to retry, recover or compensate steps that exceed their deadline or fail.

GateSift explanation

What this pattern helps you decide

Separate coordination, remote work and supervision. Durable step state and completion deadlines allow a supervisor to detect failed or timed-out work and arrange retry, recovery or compensation.

Which failure mode or capacity limit is this pattern intended to contain?
Which thresholds, deadlines, state or compensation data must remain durable?
How will operators know the pattern activated and whether recovery succeeded?
Common Azure implementations

Where you may see it

  • Durable Functions with durable orchestration state
  • Logic Apps plus durable state and recovery workflows
  • Service Bus workers coordinated through a state store and watchdog process
GateSift relevance

How the analyzers can surface it

  • Long-running workflows with no persisted progress or timeout owner
  • Background jobs that cannot resume safely after process restarts
  • Recovery logic without stable identifiers or idempotent agents

Pattern detection is contextual. GateSift should present these as architectural signals, not claim a pattern is implemented solely because one policy statement or adapter exists.

Source guidance and attribution

This page links to Microsoft Azure Architecture Center for the source architecture guidance. GateSift summaries, diagrams, Azure examples and analyzer signals are original. No endorsement by Microsoft is implied.

Back to pattern library