netflix conductor timeoutSeconds not taking effect for retries

Question

I am new to Conductor. If I understand right, when a worker task (attempt0) timeouts and gets retried according to my task definition, the timeoutSeconds should apply as well to the attempt1 and attempt2... but it looks to me it's not working this way. What am I missing here? Thanks.

My worker is sleeping for 10 minutes, attempt0 timeouts as expected, but attempt1 stays in IN_PROGRESS for 10 minutes and finally gets completed.

workflow definition

{
    "name": "timeout_test_task",
    "description": "timeout test task",
    "version": 1,
    "tasks":
    [
        {
            "name": "step_1_task",
            "taskReferenceName": "step_1_task_ref",
            "type": "SIMPLE"
        }
    ],
    "restartable": true,
    "ownerEmail": "[email protected]",
    "workflowStatusListenerEnabled": true,
    "schemaVersion": 2
}

task definition

    {
        "name": "step_1_task",
        "retryCount": 5,
        "timeoutSeconds": 10,
        "pollTimeoutSeconds": 1200,
        "timeoutPolicy": "RETRY", 
        "retryLogic": "FIXED",
        "retryDelaySeconds": 1,
        "responseTimeoutSeconds": 5,
        "concurrentExecLimit": 100,
        "rateLimitFrequencyInSeconds": 60,
        "ownerEmail": "[email protected]",
        "rateLimitPerFrequency": 2
    }

UI screenshot of attempt0

UI screenshot of attempt1

BTW, attempt0 timeouts after 31 seconds instead of 10 seconds, what's causing this delay?