7 Replies Latest reply on Feb 7, 2017 6:20 AM by cmorfin RSS
    callad Explorer

    How to use the ThingWatcher

    Dear Sir/Madam,

     

    I install the ThingAnalytics and ThingWatcher according to the installation guide. And I upload the training-service.tar, model-service.tar and install them on docker. And after that, how to use the ThingWatcher?

    Where can I get some resources, videos or something else to use the thingwatcher? How can I write the ThingWatcher?

    I want to have an example to study.

    Looking forward to your reply.

     

    Bests,

    Calla

      • Re: How to use the ThingWatcher
        ttielebein Collaborator

        There is a section on "Using ThingWatcher" in this guide, have you already seen this?

         

        Hope this helps!

          • Re: How to use the ThingWatcher
            callad Explorer

            Hi,

            I installed the docker, the Model Microservice and Training Microservice according to the ThingWatcher Development Guide.pdf. And I can get the following results.

            1.png

            Then I write a thingwathcertest1.java according to the guide.pdf, but I don't know what can i do after this? Would you mind to give me an example?

            or some more detail guide which tell me how to use the ThingWatcher?

            2 (1).png

            Thanks a lot.

            • Re: How to use the ThingWatcher
              fseheutd Newbie

              I didn't find anything about the model microservice.

               

              My understanding is that the model service is supposed to help deploying a model trained once for a device to other training microservice (eventually for other devices). But I see nowhere in the documentation, how to do so.

               

              Would you have leads on that please?

               

              Thanks in advance for your help.

                • Re: How to use the ThingWatcher
                  cmorfin Creator

                  Hi Francois Xavier

                   

                  The training microservice is the brain that does the computation and create the prediction model.

                  Once created, this model needs to be stored somewhere.

                  There are different options for that, but a convenient one is to store the model in the model microservice.

                  This then allows to restart a ThingWatcher from a previously created model directly in monitoring mode without going through calibrating mode.

                  There is a sample on how to start a ThingWatcher in monitoring mode in the Help Center.

                   

                  Hope this help

                  Kind regards

                  Christophe