在 Durable Functions (Azure Functions) 中處理錯誤

發行項
04/17/2024

Durable Function 協調流程會在程式碼中實作，且可以使用程式設計語言的內建錯誤處理功能。將錯誤處理和補償新增至協調流程時，確實已沒有任何新的概念需要了解。不過，有一些行為值得注意。

注意

適用於 Azure Functions 的第 4 版 Node.js 程式設計模型已正式推出。新的 v4 模型旨在為 JavaScript 和 TypeScript 開發人員提供更靈活的直覺式體驗。如需深入了解 v3 與 v4 之間的差異，請參閱移轉指南。

在下列程式碼片段中，JavaScript (PM4) 表示程式設計模型 V4，這是新的體驗。

活動函式中的錯誤

活動函式中擲回的任何例外狀況會封送處理回到協調器函式，並以 FunctionFailedException 擲回。您可以在協調器函式中撰寫符合需求的錯誤處理和補償程式碼。

例如，假設有下列協調器函式會從一個帳戶轉帳到另一個帳戶：

[FunctionName("TransferFunds")]
public static async Task Run([OrchestrationTrigger] IDurableOrchestrationContext context)
{
    var transferDetails = context.GetInput<TransferOperation>();

    await context.CallActivityAsync("DebitAccount",
        new
        {
            Account = transferDetails.SourceAccount,
            Amount = transferDetails.Amount
        });

    try
    {
        await context.CallActivityAsync("CreditAccount",
            new
            {
                Account = transferDetails.DestinationAccount,
                Amount = transferDetails.Amount
            });
    }
    catch (Exception)
    {
        // Refund the source account.
        // Another try/catch could be used here based on the needs of the application.
        await context.CallActivityAsync("CreditAccount",
            new
            {
                Account = transferDetails.SourceAccount,
                Amount = transferDetails.Amount
            });
    }
}

注意

先前的 C# 範例適用於 Durable Functions 2.x。針對 Durable Functions 1.x，您必須使用 DurableOrchestrationContext 而不是 IDurableOrchestrationContext。如需版本差異的詳細資訊，請參閱 Durable Functions 版本一文。

[FunctionName("TransferFunds")]
public static async Task Run(
    [OrchestrationTrigger] TaskOrchestrationContext context, TransferOperation transferDetails)
{
    await context.CallActivityAsync("DebitAccount",
        new
        {
            Account = transferDetails.SourceAccount,
            Amount = transferDetails.Amount
        });

    try
    {
        await context.CallActivityAsync("CreditAccount",
            new
            {
                Account = transferDetails.DestinationAccount,
                Amount = transferDetails.Amount
            });
    }
    catch (Exception)
    {
        // Refund the source account.
        // Another try/catch could be used here based on the needs of the application.
        await context.CallActivityAsync("CreditAccount",
            new
            {
                Account = transferDetails.SourceAccount,
                Amount = transferDetails.Amount
            });
    }
}

const df = require("durable-functions");

module.exports = df.orchestrator(function* (context) {
    const transferDetails = context.df.getInput();

    yield context.df.callActivity("DebitAccount", {
        account: transferDetails.sourceAccount,
        amount: transferDetails.amount,
    });

    try {
        yield context.df.callActivity("CreditAccount", {
            account: transferDetails.destinationAccount,
            amount: transferDetails.amount,
        });
    } catch (error) {
        // Refund the source account.
        // Another try/catch could be used here based on the needs of the application.
        yield context.df.callActivity("CreditAccount", {
            account: transferDetails.sourceAccount,
            amount: transferDetails.amount,
        });
    }
})

const df = require("durable-functions");

df.app.orchestration("transferFunds", function* (context) {
    const transferDetails = context.df.getInput();

    yield context.df.callActivity("debitAccount", {
        account: transferDetails.sourceAccount,
        amount: transferDetails.amount,
    });

    try {
        yield context.df.callActivity("creditAccount", {
            account: transferDetails.destinationAccount,
            amount: transferDetails.amount,
        });
    } catch (error) {
        // Refund the source account.
        // Another try/catch could be used here based on the needs of the application.
        yield context.df.callActivity("creditAccount", {
            account: transferDetails.sourceAccount,
            amount: transferDetails.amount,
        });
    }
});

import azure.functions as func
import azure.durable_functions as df

def orchestrator_function(context: df.DurableOrchestrationContext):
    transfer_details = context.get_input()

    yield context.call_activity('DebitAccount', {
         'account': transfer_details['sourceAccount'],
         'amount' : transfer_details['amount']
    })

    try:
        yield context.call_activity('CreditAccount', {
                'account': transfer_details['destinationAccount'],
                'amount': transfer_details['amount'],
            })
    except:
        yield context.call_activity('CreditAccount', {
            'account': transfer_details['sourceAccount'],
            'amount': transfer_details['amount']
        })

main = df.Orchestrator.create(orchestrator_function)

根據預設，PowerShell 中的 Cmdlet 不會引發可以使用 try/catch 區塊攔截的例外狀況。您有兩個選項可變更此行為：

叫用 Cmdlet 時，請使用 -ErrorAction Stop 旗標，例如 Invoke-DurableActivity。
在叫用 Cmdlet 之前，將 $ErrorActionPreference 協調器函式中的偏好設定變數設定為 "Stop"。

param($Context)

$ErrorActionPreference = "Stop"

$transferDetails = $Context.Input

Invoke-DurableActivity -FunctionName 'DebitAccount' -Input @{ account = transferDetails.sourceAccount; amount = transferDetails.amount }

try {
    Invoke-DurableActivity -FunctionName 'CreditAccount' -Input @{ account = transferDetails.destinationAccount; amount = transferDetails.amount }
} catch {
    Invoke-DurableActivity -FunctionName 'CreditAccount' -Input @{ account = transferDetails.sourceAccount; amount = transferDetails.amount }
}

如需 PowerShell 中錯誤處理的詳細資訊，請參閱 Try-Catch-Finally PowerShell 文件。

@FunctionName("TransferFunds")
public void transferFunds(
        @DurableOrchestrationTrigger(name = "ctx") TaskOrchestrationContext ctx) {
    TransferOperation transfer = ctx.getInput(TransferOperation.class);
    ctx.callActivity(
        "DebitAccount", 
        new OperationArgs(transfer.sourceAccount, transfer.amount)).await();
    try {
        ctx.callActivity(
            "CreditAccount", 
            new OperationArgs(transfer.destinationAccount, transfer.amount)).await();
    } catch (TaskFailedException ex) {
        // Refund the source account on failure
        ctx.callActivity(
            "CreditAccount", 
            new OperationArgs(transfer.sourceAccount, transfer.amount)).await();
    }
}

如果首次 CreditAccount 函式呼叫失敗，協調器函式會將款項匯回到來源帳戶，以進行彌補。

失敗時自動重試

當您呼叫活動函式或子協調流程函式時，您可以指定自動重試原則。下列範例會嘗試呼叫函式最多 3 次，並於每次重試之間等待 5 秒鐘：

[FunctionName("TimerOrchestratorWithRetry")]
public static async Task Run([OrchestrationTrigger] IDurableOrchestrationContext context)
{
    var retryOptions = new RetryOptions(
        firstRetryInterval: TimeSpan.FromSeconds(5),
        maxNumberOfAttempts: 3);

    await context.CallActivityWithRetryAsync("FlakyFunction", retryOptions, null);

    // ...
}

注意

[FunctionName("TimerOrchestratorWithRetry")]
public static async Task Run([OrchestrationTrigger] TaskOrchestrationContext context)
{
    var options = TaskOptions.FromRetryPolicy(new RetryPolicy(
        maxNumberOfAttempts: 3,
        firstRetryInterval: TimeSpan.FromSeconds(5)));

    await context.CallActivityAsync("FlakyFunction", options: options);

    // ...
}

const df = require("durable-functions");

module.exports = df.orchestrator(function*(context) {
    const firstRetryIntervalInMilliseconds = 5000;
    const maxNumberOfAttempts = 3;

    const retryOptions = 
        new df.RetryOptions(firstRetryIntervalInMilliseconds, maxNumberOfAttempts);

    yield context.df.callActivityWithRetry("FlakyFunction", retryOptions);

    // ...
});

const df = require("durable-functions");

df.app.orchestration("callActivityWithRetry", function* (context) {
    const firstRetryIntervalInMilliseconds = 5000;
    const maxNumberOfAttempts = 3;

    const retryOptions = new df.RetryOptions(firstRetryIntervalInMilliseconds, maxNumberOfAttempts);

    yield context.df.callActivityWithRetry("flakyFunction", retryOptions);

    // ...
});

import azure.functions as func
import azure.durable_functions as df

def orchestrator_function(context: df.DurableOrchestrationContext):
    first_retry_interval_in_milliseconds = 5000
    max_number_of_attempts = 3

    retry_options = df.RetryOptions(first_retry_interval_in_milliseconds, max_number_of_attempts)

    yield context.call_activity_with_retry('FlakyFunction', retry_options)

main = df.Orchestrator.create(orchestrator_function)

param($Context)

$retryOptions = New-DurableRetryOptions `
                    -FirstRetryInterval (New-TimeSpan -Seconds 5) `
                    -MaxNumberOfAttempts 3

Invoke-DurableActivity -FunctionName 'FlakyFunction' -RetryOptions $retryOptions

@FunctionName("TimerOrchestratorWithRetry")
public void timerOrchestratorWithRetry(
        @DurableOrchestrationTrigger(name = "ctx") TaskOrchestrationContext ctx) {
    final int maxAttempts = 3;
    final Duration firstRetryInterval = Duration.ofSeconds(5);
    RetryPolicy policy = new RetryPolicy(maxAttempts, firstRetryInterval);
    TaskOptions options = new TaskOptions(policy);
    ctx.callActivity("FlakeyFunction", options).await();
    // ...
}

上一個範例中的活動函式呼叫會採用設定自動重試原則的參數。自訂自動重試原則有幾個選項：

嘗試次數上限：嘗試次數上限。如果設定為 1，則不會進行重試。
第一次重試間隔：第一次重試嘗試之前等候的時間量。
輪詢係數：用來決定輪詢增加速率的係數。預設值為 1。
最大重試間隔：重試嘗試之間等候的最大時間量。
重試逾時：花費在重試的最大時間量。預設行為是無限期地重試。

自訂重試處理常式

使用 .NET 或 JAVA 時，您也可以選擇在程式碼中實作重試處理常式。當宣告式重試原則不夠用於表達時，這很實用。對於未支援自訂重試處理常式的語言，您仍然可以選擇使用迴圈、例外狀況處理及計時器來實作重試原則，以插入重試之間的延遲。

RetryOptions retryOptions = new RetryOptions(
    firstRetryInterval: TimeSpan.FromSeconds(5),
    maxNumberOfAttempts: int.MaxValue)
    {
        Handle = exception =>
        {
            // True to handle and try again, false to not handle and throw.
            if (exception is TaskFailedException failure)
            {
                // Exceptions from TaskActivities are always this type. Inspect the
                // inner Exception to get more details.
            }

            return false;
        };
    }

await ctx.CallActivityWithRetryAsync("FlakeyActivity", retryOptions, null);

TaskOptions retryOptions = TaskOptions.FromRetryHandler(retryContext =>
{
    // Don't retry anything that derives from ApplicationException
    if (retryContext.LastFailure.IsCausedBy<ApplicationException>())
    {
        return false;
    }

    // Quit after N attempts
    return retryContext.LastAttemptNumber < 3;
});

try
{
    await ctx.CallActivityAsync("FlakeyActivity", options: retryOptions);
}
catch (TaskFailedException)
{
    // Case when the retry handler returns false...
}

RetryHandler retryHandler = retryCtx -> {
    // Don't retry anything that derives from RuntimeException
    if (retryCtx.getLastFailure().isCausedBy(RuntimeException.class)) {
        return false;
    }

    // Quit after N attempts
    return retryCtx.getLastAttemptNumber() < 3;
};

TaskOptions options = new TaskOptions(retryHandler);
try {
    ctx.callActivity("FlakeyActivity", options).await();
} catch (TaskFailedException ex) {
    // Case when the retry handler returns false...
}

函式逾時

如果協調器函式內的函式呼叫太久才完成，您可以放棄此函式呼叫。目前，適當的作法是使用 "any" 工作選取器來建立永久性計時器，如下列範例所示：

[FunctionName("TimerOrchestrator")]
public static async Task<bool> Run([OrchestrationTrigger] IDurableOrchestrationContext context)
{
    TimeSpan timeout = TimeSpan.FromSeconds(30);
    DateTime deadline = context.CurrentUtcDateTime.Add(timeout);

    using (var cts = new CancellationTokenSource())
    {
        Task activityTask = context.CallActivityAsync("FlakyFunction");
        Task timeoutTask = context.CreateTimer(deadline, cts.Token);

        Task winner = await Task.WhenAny(activityTask, timeoutTask);
        if (winner == activityTask)
        {
            // success case
            cts.Cancel();
            return true;
        }
        else
        {
            // timeout case
            return false;
        }
    }
}

注意

[Function("TimerOrchestrator")]
public static async Task<bool> Run([OrchestrationTrigger] TaskOrchestrationContext context)
{
    TimeSpan timeout = TimeSpan.FromSeconds(30);
    DateTime deadline = context.CurrentUtcDateTime.Add(timeout);

    using (var cts = new CancellationTokenSource())
    {
        Task activityTask = context.CallActivityAsync("FlakyFunction");
        Task timeoutTask = context.CreateTimer(deadline, cts.Token);

        Task winner = await Task.WhenAny(activityTask, timeoutTask);
        if (winner == activityTask)
        {
            // success case
            cts.Cancel();
            return true;
        }
        else
        {
            // timeout case
            return false;
        }
    }
}

const df = require("durable-functions");
const moment = require("moment");

module.exports = df.orchestrator(function*(context) {
    const deadline = moment.utc(context.df.currentUtcDateTime).add(30, "s");

    const activityTask = context.df.callActivity("FlakyFunction");
    const timeoutTask = context.df.createTimer(deadline.toDate());

    const winner = yield context.df.Task.any([activityTask, timeoutTask]);
    if (winner === activityTask) {
        // success case
        timeoutTask.cancel();
        return true;
    } else {
        // timeout case
        return false;
    }
});

const df = require("durable-functions");
const { DateTime } = require("luxon");

df.app.orchestration("timerOrchestrator", function* (context) {
    const deadline = DateTime.fromJSDate(context.df.currentUtcDateTime).plus({ seconds: 30 });

    const activityTask = context.df.callActivity("flakyFunction");
    const timeoutTask = context.df.createTimer(deadline.toJSDate());

    const winner = yield context.df.Task.any([activityTask, timeoutTask]);
    if (winner === activityTask) {
        // success case
        timeoutTask.cancel();
        return true;
    } else {
        // timeout case
        return false;
    }
});

import azure.functions as func
import azure.durable_functions as df
from datetime import datetime, timedelta

def orchestrator_function(context: df.DurableOrchestrationContext):
    deadline = context.current_utc_datetime + timedelta(seconds = 30)
    
    activity_task = context.call_activity('FlakyFunction')
    timeout_task = context.create_timer(deadline)

    winner = yield context.task_any(activity_task, timeout_task)
    if winner == activity_task:
        timeout_task.cancel()
        return True
    else:
        return False

main = df.Orchestrator.create(orchestrator_function)

param($Context)

$expiryTime = New-TimeSpan -Seconds 30

$activityTask = Invoke-DurableActivity -FunctionName 'FlakyFunction'-NoWait
$timerTask = Start-DurableTimer -Duration $expiryTime -NoWait

$winner = Wait-DurableTask -Task @($activityTask, $timerTask) -NoWait

if ($winner -eq $activityTask) {
    Stop-DurableTimerTask -Task $timerTask
    return $True
}
else {
    return $False
}

@FunctionName("TimerOrchestrator")
public boolean timerOrchestrator(
        @DurableOrchestrationTrigger(name = "ctx") TaskOrchestrationContext ctx) {
    Task<Void> activityTask = ctx.callActivity("SlowFunction");
    Task<Void> timeoutTask = ctx.createTimer(Duration.ofMinutes(30));

    Task<?> winner = ctx.anyOf(activityTask, timeoutTask).await();
    if (winner == activityTask) {
        // success case
        return true;
    } else {
        // timeout case
        return false;
    }
}

注意

這項機制並不會實際終止進行中的活動函式執行。只是讓協調器函式略過結果並繼續執行。如需詳細資訊，請參閱計時器文件。

未處理的例外狀況

如果協調器函式失敗並傳回未處理的例外狀況，例外狀況的詳細資料會記錄下來，而在執行個體會以 Failed 狀態結束。

下一步

了解 Eternal 協調流程

了解如何診斷問題

共用方式為