Durable Functions orchestration patterns in Azure - Step-by-Step Execution

Process Flow - Durable Functions orchestration patterns

Start Orchestration

↓

Call Activity Function

↓

Wait for Activity Completion

↓

Decide Next Step

↓

Call Another

The orchestration starts, calls activity functions, waits for their results, decides next steps, and either continues or ends.

Execution Sample

Azure

1. Start orchestration
2. Call activity A
3. Wait for A to complete
4. If success, call activity B
5. End orchestration

This orchestration calls two activities in sequence, waiting for each to finish before continuing.

Process Table

Step	Action	Function Called	Status	Next Step
1	Start orchestration	Orchestrator	Running	Call activity A
2	Call activity	Activity A	Started	Wait for completion
3	Wait for completion	Activity A	Completed	Check result
4	Check result	Orchestrator	Success	Call activity B
5	Call activity	Activity B	Started	Wait for completion
6	Wait for completion	Activity B	Completed	End orchestration
7	End orchestration	Orchestrator	Completed	Stop

💡 Orchestration ends after all activities complete successfully.

Status Tracker

Variable	Start	After Step 3	After Step 6	Final
orchestrationStatus	Not started	Running	Running	Completed
activityAStatus	Not started	Completed	Completed	Completed
activityBStatus	Not started	Not started	Completed	Completed

Key Moments - 3 Insights

Why does the orchestration wait after calling an activity function?

Because Durable Functions orchestrator pauses until the activity function completes, ensuring reliable sequencing as shown in steps 2 and 3 of the execution_table.

What happens if an activity function fails?

The orchestration can catch the failure and decide next steps, but in this example, success is assumed at step 4 to continue calling the next activity.

Can orchestration call multiple activities in parallel?

Yes, but this example shows sequential calls. Parallel calls would involve starting multiple activities before waiting, changing the flow after step 2.

Visual Quiz - 3 Questions

Look at the execution_table, what is the orchestrationStatus after step 3?

A Not started

B Running

C Completed

D Failed

💡 Hint

Check the variable_tracker column 'After Step 3' for orchestrationStatus.

At which step does the orchestration decide to call activity B?

A Step 2

B Step 3

C Step 4

D Step 5

💡 Hint

Look at the 'Next Step' column in execution_table at step 4.

If activity A failed, how would the orchestration flow change?

A It would handle failure and decide next steps

B It would end immediately

C It would call activity B anyway

D It would restart activity A automatically

💡 Hint

Refer to key_moments about failure handling in orchestration.

Concept Snapshot

Durable Functions orchestrate workflows by calling activity functions.
They wait for each activity to complete before continuing.
Orchestrations can run activities sequentially or in parallel.
Failures can be caught and handled to control flow.
This ensures reliable, stateful serverless workflows.

Full Transcript

Durable Functions orchestration patterns involve starting an orchestration that calls activity functions. The orchestration waits for each activity to complete before deciding the next step. This can be sequential or parallel. The orchestration tracks status variables for each activity and itself. If an activity fails, the orchestration can catch the failure and decide how to proceed. This pattern ensures reliable and stateful workflows in serverless environments.

Practice

1. What is the main role of an orchestrator function in Azure Durable Functions?

easy

A. To perform the actual work like processing data

B. To coordinate and manage the workflow of multiple tasks

C. To store data permanently in the cloud

D. To send notifications to users

Solution

1. Step 1: Understand the function types in Durable Functions

   Durable Functions use orchestrator functions to manage workflows and activity functions to perform tasks.

2. Step 2: Identify the role of the orchestrator function

   The orchestrator function controls the order and timing of tasks but does not do the actual work itself.

3. Final Answer:

   To coordinate and manage the workflow of multiple tasks -> Option B

4. Quick Check:

   Orchestrator = workflow manager [OK]

Hint: Orchestrator controls flow; activity does the work [OK]

Common Mistakes:

• Confusing orchestrator with activity function
• Thinking orchestrator stores data
• Assuming orchestrator sends notifications

2. Which of the following is the correct way to call an activity function named ProcessOrder from an orchestrator function in C#?

easy

A. await context.CallActivityAsync("ProcessOrder", orderId);

B. context.CallActivity("ProcessOrder", orderId);

C. await CallActivityAsync("ProcessOrder", orderId);

D. context.CallActivityAsync("ProcessOrder");

Solution

1. Step 1: Recall the syntax for calling activity functions

   In Durable Functions, the orchestrator calls activities using await context.CallActivityAsync with the function name and input.

2. Step 2: Check each option for correctness

   await context.CallActivityAsync("ProcessOrder", orderId); uses the correct method with await, context, function name, and input parameter.

3. Final Answer:

   await context.CallActivityAsync("ProcessOrder", orderId); -> Option A

4. Quick Check:

   Correct async call syntax = await context.CallActivityAsync("ProcessOrder", orderId); [OK]

Hint: Use await with context.CallActivityAsync and function name [OK]

Common Mistakes:

• Omitting await keyword
• Using wrong method name like CallActivity
• Missing input parameter when required

3. Given this orchestrator code snippet in JavaScript:

const outputs = [];
outputs.push(await context.callActivity('TaskA', 1));
outputs.push(await context.callActivity('TaskB', 2));
return outputs;

What will the orchestrator return?

medium

A. An array with results from TaskA and TaskB in order

B. A single result from TaskB only

C. An empty array

D. A promise object instead of results

Solution

1. Step 1: Analyze the code execution flow

   The orchestrator calls TaskA and waits for its result, then calls TaskB and waits for its result, pushing both into the outputs array.

2. Step 2: Understand the return value

   Since both calls are awaited, outputs will contain the results of TaskA and TaskB in order.

3. Final Answer:

   An array with results from TaskA and TaskB in order -> Option A

4. Quick Check:

   Awaited calls return results in array [OK]

Hint: Await each call to get results in order [OK]

Common Mistakes:

• Assuming only last result is returned
• Thinking outputs is empty without awaits
• Confusing promise with resolved value

4. You wrote this orchestrator function in C#:

public async Task<string> RunOrchestrator(IDurableOrchestrationContext context)
{
    var result = context.CallActivityAsync<string>("DoWork", null);
    return result.Result;
}

What is the problem with this code?

medium

A. It calls the wrong method for activity

B. It correctly returns the activity result

C. It misses the await keyword causing a compile error

D. It blocks the orchestrator causing a deadlock

Solution

1. Step 1: Identify async call usage

   The code calls CallActivityAsync but does not await it, instead accesses result.Result synchronously.

2. Step 2: Understand deadlock risk in orchestrators

   Accessing Result blocks the thread and can cause deadlocks in async orchestrator functions.

3. Final Answer:

   It blocks the orchestrator causing a deadlock -> Option D

4. Quick Check:

   Use await, not .Result, to avoid deadlocks [OK]

Hint: Always await async calls in orchestrators [OK]

Common Mistakes:

• Using .Result instead of await
• Ignoring async method patterns
• Assuming synchronous access works fine

5. You want to run three activity functions Task1, Task2, and Task3 in parallel and wait for all to finish before continuing. Which orchestrator pattern correctly achieves this in JavaScript Durable Functions?

hard

A. await context.callActivity('Task1');
await context.callActivity('Task2');
await context.callActivity('Task3');

B. const results = [];
results.push(await context.callActivity('Task1'));
results.push(await context.callActivity('Task2'));
results.push(await context.callActivity('Task3'));

C. const tasks = [
context.callActivity('Task1'),
context.callActivity('Task2'),
context.callActivity('Task3')
];
const results = await Promise.all(tasks);

D. const results = await context.callActivity('Task1') + await context.callActivity('Task2') + await context.callActivity('Task3');

Solution

1. Step 1: Understand parallel execution in JavaScript

   To run tasks in parallel, start them without awaiting immediately, collect promises, then await all together.

2. Step 2: Analyze each option

   const tasks = [ context.callActivity('Task1'), context.callActivity('Task2'), context.callActivity('Task3') ]; const results = await Promise.all(tasks); creates an array of promises and awaits them all with Promise.all, running tasks concurrently.

3. Final Answer:

   Use Promise.all with array of activity calls for parallel execution -> Option C

4. Quick Check:

   Parallel = start all, then await all [OK]

Hint: Use Promise.all to await multiple tasks in parallel [OK]

Common Mistakes:

• Awaiting each task sequentially (Options A and C)
• Trying to add awaited results (Option D)
• Not collecting promises before awaiting