Default runtime timeout

[nadeshseen]

is it possible to change default runtime timeout without creating a custom runtime using FunctionExecutor()?
Similar to fexec = lithops.FunctionExecutor(runtime_memory=256)

[JosepSampe]

Any parameter that can be set in the openwhisk section of the config file, can be set in the FunctionExecutor().
So you can do this: fexec = lithops.FunctionExecutor(runtime_timeout=300)

[nadeshseen]

But, fexec = lithops.FunctionExecutor(runtime_timeout=1000) this doesn't seem to work. Default timeout remains 300s.

ok: got activation f9d12e70977a45b8912e70977a25b8e2
{
    "namespace": "guest",
    "name": "docker.io_lithopscloud_openwhisk-python-v310_256MB_3.1.1.dev0",
    "version": "0.0.1",
    "subject": "guest",
    "activationId": "f9d12e70977a45b8912e70977a25b8e2",
    "start": 1707993352282,
    "end": 1707993362568,
    "duration": 10286,
    "statusCode": 0,
    "response": {
        "status": "success",
        "statusCode": 0,
        "success": true,
        "result": {
            "Execution": "Finished"
        }
    },
    "logs": [
        "2024-02-15T10:35:52.303179448Z stdout: [INFO] __main__.py:43 -- Lithops v3.1.1.dev0 - Starting OpenWhisk execution",
        "2024-02-15T10:35:52.32074551Z  stdout: [INFO] minio.py:61 -- MinIO client created - Endpoint: http://10.129.28.254:31600",
        "2024-02-15T10:35:52.35844187Z  stdout: 2024-02-15 10:35:52,357 [INFO] handler.py:74 -- Tasks received: 1 - Worker processes: 1",
        "2024-02-15T10:35:52.358618255Z stdout: 2024-02-15 10:35:52,358 [INFO] handler.py:116 -- Worker process 0 started",
        "2024-02-15T10:35:52.361414203Z stdout: 2024-02-15 10:35:52,361 [INFO] handler.py:176 -- Lithops v3.1.1.dev0 - Starting OpenWhisk execution",
        "2024-02-15T10:35:52.361533786Z stdout: 2024-02-15 10:35:52,361 [INFO] handler.py:177 -- Execution ID: c968c3-0-M000/00000",
        "2024-02-15T10:35:52.370751587Z stdout: 2024-02-15 10:35:52,370 [INFO] minio.py:61 -- MinIO client created - Endpoint: http://10.129.28.254:31600",
        "2024-02-15T10:35:52.420220262Z stdout: 2024-02-15 10:35:52,419 [INFO] jobrunner.py:234 -- Going to execute 'my_map_function()'",
        "2024-02-15T10:35:52.420342424Z stdout: ---------------------- FUNCTION LOG ----------------------",
        "2024-02-15T10:35:52.420367465Z stdout: I'm activation number 0",
        "2024-02-15T10:35:52.420375928Z stdout: ----------------------------------------------------------",
        "2024-02-15T10:35:52.42059047Z  stdout: 2024-02-15 10:35:52,420 [INFO] jobrunner.py:240 -- Success function execution",
        "2024-02-15T10:35:52.424888872Z stdout: 2024-02-15 10:35:52,424 [INFO] jobrunner.py:312 -- Process finished",
        "2024-02-15T10:35:52.432255146Z stdout: 2024-02-15 10:35:52,431 [INFO] status.py:88 -- Storing execution stats - Size: 1.1KiB",
        "2024-02-15T10:36:02.555235972Z stdout: 2024-02-15 10:36:02,554 [INFO] handler.py:270 -- Finished",
        "2024-02-15T10:36:02.555650611Z stdout: 2024-02-15 10:36:02,555 [INFO] handler.py:138 -- Worker process 0 finished"
    ],
    "annotations": [
        {
            "key": "path",
            "value": "guest/lithops_23bc4/docker.io_lithopscloud_openwhisk-python-v310_256MB_3.1.1.dev0"
        },
        {
            "key": "waitTime",
            "value": 40
        },
        {
            "key": "kind",
            "value": "blackbox"
        },
        {
            "key": "timeout",
            "value": false
        },
        {
            "key": "limits",
            "value": {
                "concurrency": 1,
                "logs": 10,
                "memory": 256,
                "timeout": 300000
            }
        }
    ],
    "publish": false
}

[JosepSampe]

runtime_timeout is applied at the deployment phase. You have to remove your previously deployed function if you want this parameter to take effect.

If what you want is to set a different timeout for each different function invocation, it is convenient to set the runtime_timeout parameter in your config file to a high value, and then use the execution_timeout parameter (< runtime_timeout) of the fexec.map() call.