Map AggegateError and Error.cause to ApplicationFailureInfo

[mdouglass]

What was changed

Change how ensureApplicationFailure and the DefaultFailureConverter handle Error.cause and AggregateError.errors.

Why?

• ensureApplicationFailure was not providing Error.cause to ApplicationFailure.create which meant that cause information was not being built

• AggegateError.errors wasn't being handled in either ensureApplicationFailure or the DefaultFailureConverter. It is not possible to map that directly to an IFailure, so that information is added to then ApplicationFailureInfo.details field.

Checklist

1. Closes [Feature Request] support AggregateError #1675

2. How was this tested:
   I was using an activity that looks like this:

export async function failHard( ): Promise<void> {
  throw new AggregateError(
    [
      new Error('Analyze workflow gets all the love.', {
        cause: new Error('Analyze workflow is the money machine.'),
      }),
      new AggregateError(
        [
          new Error('Everyone knows acquire does the hard work'),
          new Error('Everyone can see what acquire does'),
        ],
        'Acquire workflow gets all the praise.',
      ),
    ],
    'Scrape workflow quits!',
    {
      cause: new AggregateError(
        [
          new Error('Scrape workflow is a drama queen'),
          new Error('Did I mention, scrape workflow is a drama queen?'),
        ],
        'Scrape workflow is a drama queen x2.',
      ),
    },
  )
}

Previously, this was all the information available in the Temporal web site for the error:

{
  "type": "workflowExecutionFailedEventAttributes",
  "failure":
    {
      "message": "Activity task failed",
      "cause":
        {
          "message": "Scrape workflow quits!",
          "source": "TypeScriptSDK",
          "stackTrace": "AggregateError: Scrape workflow quits!\n    at getLastScrape (file:///Users/matthew/adology/adology-backend/packages/ad-scraper/src/activities/last-scrape.ts:16:9)",
          "applicationFailureInfo": { "type": "AggregateError" },
        },
      "activityFailureInfo":
        {
          "scheduledEventId": "5",
          "startedEventId": "6",
          "identity": "[email protected]",
          "activityType": { "name": "getLastScrape" },
          "activityId": "1",
          "retryState": "RETRY_STATE_MAXIMUM_ATTEMPTS_REACHED",
        },
    },
  "retryState": "RETRY_STATE_RETRY_POLICY_NOT_SET",
  "workflowTaskCompletedEventId": "10",
}

After this PR, this is how the failed activity is reported:

{
  "type": "workflowExecutionFailedEventAttributes",
  "failure": {
    "message": "Activity task failed",
    "cause": {
      "message": "Scrape workflow quits!",
      "source": "TypeScriptSDK",
      "stackTrace": "AggregateError: Scrape workflow quits!\n    at getLastScrape (file:///Users/matthew/adology/adology-backend/packages/ad-scraper/src/activities/last-scrape.ts:16:9)",
      "cause": {
        "message": "Scrape workflow is a drama queen x2.",
        "source": "TypeScriptSDK",
        "stackTrace": "AggregateError: Scrape workflow is a drama queen x2.\n    at getLastScrape (file:///Users/matthew/adology/adology-backend/packages/ad-scraper/src/activities/last-scrape.ts:31:14)",
        "applicationFailureInfo": {
          "details": {
            "payloads": [
              [
                {
                  "source": "TypeScriptSDK",
                  "message": "Scrape workflow is a drama queen",
                  "stackTrace": "Error: Scrape workflow is a drama queen\n    at getLastScrape (file:///Users/matthew/adology/adology-backend/packages/ad-scraper/src/activities/last-scrape.ts:33:11)"
                },
                {
                  "source": "TypeScriptSDK",
                  "message": "Did I mention, scrape workflow is a drama queen?",
                  "stackTrace": "Error: Did I mention, scrape workflow is a drama queen?\n    at getLastScrape (file:///Users/matthew/adology/adology-backend/packages/ad-scraper/src/activities/last-scrape.ts:34:11)"
                }
              ]
            ]
          }
        }
      },
      "applicationFailureInfo": {
        "type": "AggregateError",
        "details": {
          "payloads": [
            {
              "message": "Analyze workflow gets all the love.",
              "type": "Error",
              "stack": "Error: Analyze workflow gets all the love.\n    at getLastScrape (file:///Users/matthew/adology/adology-backend/packages/ad-scraper/src/activities/last-scrape.ts:18:7)",
              "cause": {
                "message": "Analyze workflow is the money machine.",
                "type": "Error",
                "stack": "Error: Analyze workflow is the money machine.\n    at getLastScrape (file:///Users/matthew/adology/adology-backend/packages/ad-scraper/src/activities/last-scrape.ts:19:16)"
              }
            },
            {
              "message": "Acquire workflow gets all the praise.",
              "type": "AggregateError",
              "stack": "AggregateError: Acquire workflow gets all the praise.\n    at getLastScrape (file:///Users/matthew/adology/adology-backend/packages/ad-scraper/src/activities/last-scrape.ts:21:7)",
              "details": [
                {
                  "message": "Everyone knows acquire does the hard work",
                  "type": "Error",
                  "stack": "Error: Everyone knows acquire does the hard work\n    at getLastScrape (file:///Users/matthew/adology/adology-backend/packages/ad-scraper/src/activities/last-scrape.ts:23:11)"
                },
                {
                  "message": "Everyone can see what acquire does",
                  "type": "Error",
                  "stack": "Error: Everyone can see what acquire does\n    at getLastScrape (file:///Users/matthew/adology/adology-backend/packages/ad-scraper/src/activities/last-scrape.ts:24:11)"
                }
              ]
            }
          ]
        }
      }
    },
    "activityFailureInfo": {
      "scheduledEventId": "5",
      "startedEventId": "6",
      "identity": "[email protected]",
      "activityType": { "name": "getLastScrape" },
      "activityId": "1",
      "retryState": "RETRY_STATE_MAXIMUM_ATTEMPTS_REACHED"
    }
  },
  "retryState": "RETRY_STATE_RETRY_POLICY_NOT_SET",
  "workflowTaskCompletedEventId": "10"
}

3. Any docs updates needed?

[CLAassistant]

All committers have signed the CLA.

[mdouglass]

This does what I'm looking for, in that it captures the Error.cause and AggregateError.errors fields, but I don't know that it's perfect.

1. In an ideal world, the underlying IFailure's cause field would be an array so we could use that instead of dropped the aggregates into details.
2. It feels a bit weird that I have to duplicate the same logic in both ensureApplicationFailure and the DefaultFailureConverter. You probably don't want to change ensureApplicationFailure to call ApplicationFailure.create({ cause: error }) since it would add an extra layer of nesting in the error, but maybe there's a way to clean this up I'm not seeing.