[BUG] REQUEST Connector pins CPU at maximum

[lindnma]

Describe the bug
When at least one mapping is assigned (aka. at least one endpoint needs to be queried) the iot-gateway spins the CPU up to whatever the maximum is. In case of a k8s deployment this means it even runs into CPU throttling.

Connector name (If bug in the some connector):
REQUEST Connector

Versions (please complete the following information):

• OS: Docker Image: thingsboard/tb-gateway
• Version: 3.6.1
• Python version: ?

[lindnma]

I have also taken a cursory look at the code, and the issue seems to be here:
thingsboard_gateway/connectors/request/request_connector.py Line 84 to 95

def run(self):
    while not self.__stopped:
        if self.__requests_in_progress:                 # This seems to be true as soon as one mapping is
            for request in self.__requests_in_progress: # configured, causing the while loop to always run here
                if time() >= request["next_time"]:
                    thread = Thread(target=self.__send_request, args=(request, self.__convert_queue, self._log),
                                    daemon=True,
                                    name="Request to endpoint \'%s\' Thread" % (request["config"].get("url")))
                    thread.start()
        else:         # The above code causes the sleep to never execute so the code will always immediately
            sleep(.2) # rerun and never put the thread to sleep

A potential fix could be to simply remove the else before the sleep directive and always allow the thread to sleep for .2s, tho I am not deep enough in the matter to know if this would have other ramifications.

[imbeacon]

Hi @lindnma,

Thank you for your interest in ThingsBoard IoT gateway, participation in project life and your investigation. We have added a fix for this issue, it will be available in next release.