Adds tasks test for serverless

[picandocodigo]

AFAIK we can't see the list of tasks in Serverless, so I wrote a test that catches the 404 when trying to get an invalid task.

"type":"resource_not_found_exception",
"reason":"task [node_id:123456] belongs to the node [node_id] which isn't part of the cluster and there is no record of the task"

[pquentin]

Thanks! I think that test is fine, I'm just not seeing what that has to do with not being able to list tasks.

[picandocodigo]

@pquentin I mentioned listing tasks because I don't know how we can get the id of a running task from the API in Serverless to use it in task.get.

[pquentin]

Ah got it thanks. You can create a task by calling any API that has a wait_for_completion parameter, such as say the Update by query API. The task id will be in the response.

[picandocodigo]

Even with the task id, if I understand correctly Serverless won't give information on nodes, and the id needs to be node_id:task_id. So I'm not really sure how this API can be used in Serverless 🤔
I'm ok with merging the test as it is anyway.

[pquentin]

I've approved last week already, meaning I'm happy to merge this!

I did test however, and yes, this API can be used in Serverless, using those three API calls for example:

PUT /index
POST index/_update_by_query?wait_for_completion=false
GET _tasks/LO253FbxRAutdimOeNX5iw:13676458

The last API call returns this:

{
  "completed": true,
  "task": {
    "node": "serverless",
    "id": 13676458,
    "type": "transport",
    "action": "indices:data/write/update/byquery",
    "status": {
      "total": 0,
      "updated": 0,
      "created": 0,
      "deleted": 0,
      "batches": 0,
      "version_conflicts": 0,
      "noops": 0,
      "retries": {
        "bulk": 0,
        "search": 0
      },
      "throttled_millis": 0,
      "requests_per_second": -1,
      "throttled_until_millis": 0
    },
    "description": "update-by-query [index]",
    "start_time_in_millis": 1728904451130,
    "running_time_in_nanos": 3922972,
    "cancellable": true,
    "cancelled": false,
    "headers": {
      "trace.id": "a0f9f4548e8201332f63305709812b2a"
    }
  },
  "response": {
    "took": 3,
    "timed_out": false,
    "total": 0,
    "updated": 0,
    "created": 0,
    "deleted": 0,
    "batches": 0,
    "version_conflicts": 0,
    "noops": 0,
    "retries": {
      "bulk": 0,
      "search": 0
    },
    "throttled": "0s",
    "throttled_millis": 0,
    "requests_per_second": -1,
    "throttled_until": "0s",
    "throttled_until_millis": 0,
    "failures": []
  }
}

[picandocodigo]

Great, thanks @pquentin! I've updated both tasks tests to include this 👍

[pquentin]

This fails for me:

AssertionError: Failed to catch 'resource_not_found_exception' in {'completed': True, ...}

I'm assuming the same applies to the stack test. Should we refactor the common part as well?

[picandocodigo]

Ouch, sorry, that was me not committing the right code! I've updated it to check for a task key which according to my tests should work even if completed: false.