App Registry¶

On this page you will learn how to manage your your Kelvin Apps from the Kelvin Manager UI. There are new features coming every month, so be sure to check back regularly for updates.

Kelvin Apps¶

Kelvin Apps are sandboxed applications that can be developed and managed centrally on the Cloud Server and deployed to run at the edge.

At the edge they will be managed by the Kelvin Cluster. In the example here you can see a new Kelvin App being deployed from the App Registry.

There are three types of Kelvin Apps available;

1. Kelvin Developed Apps (such as Kelvin Bridges, etc.)
2. Custom Kelvin Python Apps (Any python application)
3. Custom Generic Docker Apps (Anything)

The Kelvin Cluster will manage the resources allocated to each Kelvin App to ensure that the computer that the Kelvin Node is installed on does not run out of cpu or memory resources.

Kelvin Bridges on the Kelvin Cluster ¶

Kelvin Bridges are actually special Kelvin Apps developed by Kelvin to communicate to different types of servers, PLC's, historians, SCADA systems, direct to sensors, etc., access the data and transmit this to the Cloud Server into the Kelvin Metrics.

There are current three types of Kelvin Bridges available;

• Kelvin OPC UA Bridge
• Kelvin MQTT Bridge
• Kelvin ROC Bridge

Click on the relevant bridge above to understand more about the protocol and instructions on how to deploy, edit and troubleshoot each bridge.

Kelvin Python Apps on the Kelvin Cluster ¶

Kelvin Python Apps are python applications that can process data and information at the edge. This allows for real-time logic and control decisions at the edge. Or you can do some data pre-processing before sending it to the Cloud Server.

Using Kelvin Python Apps rather than a Kelvin Docker App gives you better integration with the Node's hardware. Kelvin Python Apps are optimized so you can easily incorporate internal data routing, pipeline support, machine learning algorithms some data persistence options. This is all accessed through libraries in the Kelvin SDK.

This allows you to focus on the actual data processing code and leave the application processes, resources and life cycle management to the Kelvin Core Engine in the Kelvin Cluster.

To understand more you can read the documentation on Kelvin Python Apps here.

Kelvin Docker App ¶

Kelvin Docker Apps give you complete freedom to develop any type of Kelvin App in any language and structure.

It iIt is the same as creating a standard Docker Container using a Dockerfile configuration file along with any files to be incorporated into the container.

Some common Kelvin Docker App applications can be;

• Create Flask application to have local web application
• Create a Node Red instance for a local SCADA system
• Create a local MQTT broker for local communications
• Create an InfluxDB database instance for local data retention
• Create an nginx web server

To understand more you can read the documentation on the Kelvin Generic App here.

Overview¶

The App Registry is where you can view all the Kelvin Apps you can created and stored in the Cloud Server.

To access the App Registry page simply click on App Registry under Repository.

App Registry UI¶

Searching / Filtering¶

You can search the App Registry by their Display Name.

On the left hand box, you can enter any text. The view below will be automatically updated with every key you press.

The search will look for any matches of the word you type. The word does not need to start at the beginning of the Display Name. For example if you type machine, then the search will find Kelvin Apps named "Demo Machine Learning App" and also "machine learning anomaly monitor".

Sort View ¶

You can sort by ascending or descending order in any column. Simply click the column heading name.

Clicking the first time will sort in ascending order, clicking a second time will sort in descending order and if you click a third time the sort will be cancelled.

Kelvin App UI¶

You can see the details of all versions of a Kelvin App by clicking on it's name or on the eye icon.

Workloads¶

In the overview page you can see where the Kelvin App has been deployed, its current status and other miscellaneous information.

Kelvin App¶

Action	Description
	You can select which version of Kelvin App to filter. It will only show a list of Kelvin App deployments that meet the version number selected.
	There are five types of Kelvin Apps mentioned here; kelvin, docker, poller, bridge. Kelvin : A Kelvin Python App. Docker : A Kelvin Docker App developed by any client. Poller : Bridge : A special App developed by the Kelvin development team to connect to assets automatically with MQTT, OPC UA and Emerson ROC.
	This is the status of the Kelvin App in the App Registry.
	This will delete the application from all Kelvin Clusters, the Kelvin Core and all it's versions.

Delete Application Version¶

You can delete only certain versions of an application. Simply click on the Version drop down box and click on the delete button of the version you want to delete.

After clicking you will get a delete confirmation popup.

Individual Deployed App Details¶

Action	Description
	The status of the deployed Kelvin App on the Node.
	The time when the Kelvin App was first deployed to the Kelvin Cluster.
	This is the Display Name of the deployed Kelvin App. Click here will take you to the Connection -> Nodes -> App Details page. See here for more details.
	This is the Kelvin Cluster where the Kelvin App is deployed and running on. Click here will take you to the Connection -> Nodes Details page. See here for more details.
	This is the Name of the deployed Kelvin App.
	This is how long since the App was last detected online.

Actions¶

Action	Description
	This button will show you the Kelvin App logs. More details about this page can be found here.
	This button will control the Kelvin App. Depending on its status you can see Stop, Start or Reset button.
	This menu item in more button will allow you to edit the app.yaml file for the app. More details about this page can be found here.
	This menu item in more button will delete the selected Kelvin App on the selected Node. This will not delete the Kelvin App in the App Repository on the Cloud Server. You will be asked to confirm before deleting.

Details¶

This shows the app.yaml file which is written when creating a new core app (Kelvin App). You can view it in YAML or JSON format.

Create Flow App¶

This feature is now discontinued and is available only for backwards compatibility.

Create Core App (Kelvin App)¶

At the moment you can only create the Kelvin Python App and Kelvin Docker App with the command line SDK tools. Check back regularly as we add new features to the Kelvin Manager UI monthly and this will be included in future versions.

You can read the instructions by clicking on the Create Core App.

The instructions on how to create a core app (Kelvin App) show in a popup;

Creating a Kelvin App¶

Kelvin Bridge App¶

There are current three types of Kelvin Bridges available;

• Kelvin OPC UA Bridge
• Kelvin MQTT Bridge
• Kelvin ROC Bridge

Click on the relevant bridge above to understand more about the protocol and instructions on how to deploy, edit and troubleshoot each bridge.

Python or Docker Kelvin App¶

At the moment Python and Docker Kelvin Apps can only be created using the Kelvin SDK at the command line.

For Kelvin Python Apps you can following this link to understand how to create, handle of data and the other features available.

For Kelvin Docker Apps (also know as Generic Apps) you can follow this link to understand how to create one.

What is App Resource Allocation ?¶

The amount of CPU and memory reserved for each app can be individually tailored to suit the app's requirements. Complex Kelvin Apps will require more and simpler ones will require less.

By managing these resources manually, you can maximize how many Kelvin Apps can run on your existing Kelvin Cluster hardware.

Types of Resource Allocations¶

Limits¶

This is a maximum resource limit enforced by Kelvin Cluster. Each Kelvin App will not be allowed to use more than the limit set.

Requests¶

This is the minimum resources that is allocated to the Kelvin App. This is reserved for the Kelvin App and can not be used by other Kelvin Apps.

If there are extra resources available, the Kelvin App can use more than the requested resources as long as it does not exceed the Limits.

Resource Definitions¶

CPU¶

CPU is allocated in absolute quantities only. 1.0 CPU is equivalent to one hyperthread on an Intel core processor.

There are two different ways that you can write the allocation. For example 1/10 of one hyperthread is;

• 0.1
• 100m

Memory¶

You can allocate memory as either an integer in bytes or in multiples of G, M, k or Gi, Mi, Ki. When defining in the app.yaml do not write the "B" for bytes.

For example the number below is approximately the same;

• 128974848
• 129e6
• 129M
• 123Mi

App Resource Allocation¶

Automatic / First Deploy¶

When you first launch a Kelvin App, the Kelvin Cluster will automatically set the allocations for the request and limit resources. The default settings are;

• CPU : 0.1 or 100m
• Memory : 256Mi

Custom Allocation¶

At any time you can change the settings of the allocations for the request and limit resources.

You can do this in the Control Manage UI. First go to Connections -> Nodes

Then select the Kelvin Cluster that you want to modify the App Resource Allocations.

Now select the App that you want to change the resource allocations.

You can see the current resource allocation in the top right hand corner. If they are not set then wait a few minutes for the refresh to take effect.

Now click on the configuration tab

And you can add the following yaml settings at the bottom and put in your custom settings (If the yaml is already there, then you only need to change the settings);

  resources:
    limits:
      cpu: 100m
      memory: 512Mi
    requests:
      cpu: 100m
      memory: 512Mi

You can also edit in JSON if you are more comfortable with this format;

        "resources": {
            "limits": {
                "cpu": "100m",
                "memory": "512Mi"
            },
            "requests": {
                "cpu": "100m",
                "memory": "512Mi"
            }
        }

When you are happy with the new App Resource Allocations the you can click on Save & Deploy

Congratulations, you have created a custom app dynamic resource allocation.

Troubleshooting & Tips¶

Other Related Links¶

• Kelvin Clusters
• Kelvin SDK

Last Modified¶

Kelvin Documentation AI Support

Hi. My name is KevDocBot. How can I help you?

Send