270: Exploring the Kelvin UI for Platform Administrators¶

Where this course fits¶

This is the introductory course for the Platform Administrator path. It sits before the hands-on setup courses.

1. This course — what the Administration area contains and what each section does
2. 271: Setting Up Platform Infrastructure — clusters, nodes, and user setup
3. 272: Setting Up Assets and Data Streams Quickly — bulk asset and data stream creation

The Administration Area at a Glance¶

When you log in to Kelvin with Platform Administrator credentials, you will see two main sections in the left sidebar: Operations and Administration.

Operations is where your Engineering team works. Administration is yours.

The Administration area gives you control over nine distinct areas of the platform.

Section 1: Dashboard¶

The Administration Dashboard is the first thing you see when you enter the Administration area. It gives you a platform-wide health snapshot without having to dig into individual pages.

The dashboard is fixed — you cannot customise its layout — but you can filter it to focus on what matters to you.

Filters available:

When you apply multiple filters together, the platform shows only statistics that meet all the selected criteria.

What the dashboard shows:

The statistics are grouped into four themes.

Use the dashboard to spot trends: a sudden drop in active assets, an unusual spike in control changes, or a cluster running an old version.

For full documentation, visit:

• Administration Dashboard

Section 2: Asset Management¶

Assets are the industrial equipment your platform monitors and controls. Before Operations Engineers can do anything with an asset in the Kelvin UI, you need to create it here.

The asset model has a clear hierarchy. Understanding it before you start prevents most setup errors.

Asset Types¶

An Asset Type is a category of equipment. Every asset must belong to a type. You create types first.

Examples: beam_pump, gas_compressor, wind_turbine.

From the Administration area you can:

• Create and edit Asset Types
• Optionally configure Status Indicators — rules that change a visual indicator on the asset based on live data stream values

Assets¶

An Asset is an individual piece of equipment. It references an Asset Type, has a display name and a Name ID, and can hold optional metadata.

From the Administration area you can:

• Create assets one by one
• Bulk import many assets at once via CSV — the fastest path for large rollouts
• Edit and delete assets

Optional fields when creating an asset:

Asset Properties¶

Custom Properties¶

Asset Properties are custom key/value metadata you attach to an asset. They describe the physical characteristics of the equipment.

Properties are freeform — you define the keys and values that make sense for your operation. Common examples: Model, Manufacturer, Serial Number, Site, PLC Type.

Properties can be read by SmartApps, which allows programs to behave differently depending on the asset's attributes.

Location Properties¶

latitude and longitude properties are reserved property names.

THey can be set on each asset and set where the asset will appear on the Asset Maps view, giving Operations Engineers a geographic overview of the fleet. This is set when creating or editing an asset.

For full documentation, visit:

• Create Asset Type
• Create Asset
• Bulk Create Assets

Section 3: Data Streams¶

A Data Stream is a named signal associated with an asset. It represents one channel of data — a measurement coming in from a sensor, a set point being sent out to a controller, or a computed value from a SmartApp.

You define data streams in Administration. Connections then link live data from your assets to those streams.

Four stream types¶

Semantic types and units¶

When you create a data stream, you can (and should) assign a Semantic Type and a Unit.

A Semantic Type gives the stream meaning beyond its raw format. Kelvin includes over 40 built-in semantic types: pressure, temperature, velocity, voltage, and many more. Developers can also define custom semantic types.

A Unit defines the engineering unit for the value: bar, celsius, rpm, volt, and many more. Units can also be created on the fly during stream creation if the unit you need does not exist yet.

This metadata is used by SmartApps to understand what they are working with, and by Operations Engineers to see correctly labelled values in the UI.

For full documentation, visit:

• Create Data Stream
• Semantics and Units Reference
• Bulk Create Data Streams

Section 4: Data Quality¶

Once data is flowing from your assets, how do you know if it is actually good? That is what the Data Quality section answers.

Kelvin runs continuous quality checks on your incoming data and scores each asset/stream pair every 5 minutes.

The Data Quality Dashboard¶

The dashboard gives you an immediate visual read on the health of your data across the platform.

• The top of the dashboard shows platform-wide statistics: the percentage of good vs bad data over time
• The asset list below shows individual per-asset scores with a column for each metric

You can filter by Asset Type, individual Asset, and time window (1 hour to 7 days).

The five built-in metrics¶

Each metric is enabled or disabled per asset/stream pair. If a metric is not enabled for a pair, the dashboard shows N/A for that cell.

Drilling into a problem¶

If you spot a bad-data flag on an asset, click the analytics icon next to that asset. You can then select a data stream to see detailed metric results on a timeseries graph. From there, a link takes you directly to the Connection workload logs to help you trace the source of the problem.

For full documentation, visit:

• Data Quality Dashboard

Section 5: Guardrails¶

Guardrails are the final layer of protection before any value is written to an asset. They apply to all writes — whether the instruction comes from the UI, the API, or a SmartApp.

Think of them as a hard stop that no program or user can bypass.

What guardrails protect against¶

When a guardrail rule is breached, the control change automatically fails. The original value remains on the asset. An annotation is added to the control change record, which you can see in the Data Explorer.

The three rule types¶

There are three ways to set a guardrail on an asset/stream pair:

You can combine all three on the same asset/stream pair.

Static vs dynamic thresholds¶

Each limit can be defined as:

• A static value — a fixed number that never changes
• A dynamic data stream — the live value of another data stream on the platform, allowing limits to adapt to current operating conditions

For full documentation, visit:

• Guardrails Overview
• Create Guardrail

Section 6: Cluster Management¶

This section is where you manage the edge infrastructure you set up in course 271. Once clusters are running, you return here to monitor them and grow them.

What you can do¶

Monitoring cluster health¶

The cluster view shows you the Kelvin version running on each cluster. This matters when updates are released — you can see at a glance which clusters are current and which need attention.

For full documentation, visit:

• Cluster Management
• Add a Node
• Kelvin Edge Services

Section 7: Connections¶

A Connection is a deployed workload on a cluster that collects data from your assets and writes it to Kelvin data streams — or sends instructions back to assets.

Connections are built by Developers and uploaded to the platform. As an Administrator, your job is to deploy them, configure them, and keep them running.

What you manage in Connections¶

IO Maps are the core of a connection. An IO Map defines which asset is linked to which data stream. When you deploy a connection, you configure its IO map to match your assets.

Key tasks¶

For full documentation, visit:

• Update a Connection
• Delete a Connection

Section 8: User Management¶

This section is covered in detail in course 271. Here is a summary of everything available to you.

User management in Kelvin controls who can access the Kelvin Platform — not clusters, not SmartApps directly, but the Kelvin UI and API. Once a user is authenticated, their RBAC roles and ABAC policies determine what they can see and do.

What you manage here¶

Users — Create, edit, and delete individual Platform users. Set temporary passwords, force password changes on first login, and require 2FA.

Groups — Create groups with a standard set of roles. Assign users to groups instead of assigning roles individually. This is the recommended approach for any deployment with more than a handful of users.

Roles and permissions (RBAC) — Assign roles at the group or user level. Roles define what actions (create, read, update, delete) a user can perform on what resource types (assets, data streams, applications, etc.).

Attribute policies (ABAC) — Write YAML-based policies that filter access based on resource attributes. For example, restrict a group to only see assets belonging to a specific asset type.

For full documentation, visit:

• Manage Users
• Define User Groups
• RBAC
• ABAC

Section 9: Telemetry and Logs¶

When things are running, you need to know they are actually working. Kelvin gives you visibility into the health of every workload and connection on the platform through built-in Logs and Telemetry tabs, plus a centralised Grafana dashboard for a platform-wide view.

You can access these tools in two places:

• Directly from the Applications and Connections pages for per-workload or per-connection detail
• From the Monitoring menu in the sidebar for a platform-wide view via Grafana

Workload Status¶

Before checking logs, start with the workload's current state. Click on Applications, select the SmartApp card, and go to the Workload tab.

Each workload shows one of several states:

If a workload has a staged version waiting, you can expand the row to see its details before activating it.

Workload Logs¶

From the Workload tab, click on a workload name to open its detail page, then click the Logs tab.

Logs show everything the workload is outputting: errors, warnings, status messages, and any custom output from the SmartApp code.

There are two modes for viewing logs:

Use Live Data when you have just deployed a workload and want to see what it is doing right now. Use Forwarded Data when you want to review a time range of historical output.

Workload Telemetry¶

From the same workload detail page, click the Telemetry tab. This shows the compute health of the workload — not what it is doing logically, but how much resource it is consuming.

Metrics available:

• Memory Usage (Requests and Limits)
• CPU Usage (Requests and Limits)
• Network Usage (Received and Transmitted)
• Storage IO
• IOPS and Throughput

Use telemetry to spot workloads that are running out of memory, consuming too much CPU, or generating unexpected network traffic.

Connection Logs and Telemetry¶

Connections have the same Logs and Telemetry tabs available. Click on Connections, select a connection, and navigate to either tab.

The same Live/Forwarded Data toggle applies to connection logs. Telemetry metrics are identical to those for workloads.

Grafana Dashboards (Platform-Wide View)¶

For a platform-wide view across all clusters, nodes, workloads, and connections, Kelvin provides a centralised Grafana dashboard.

Access it by clicking Monitoring in the left sidebar.

Grafana connects to four data sources:

You can use Grafana to build custom dashboards and set up alerts — for example, a notification when any workload restarts or when a node is running low on memory. This is particularly useful for proactive platform management.

For full documentation, visit:

• Monitor Dashboard
• Grafana Dashboards
• View Workload Status
• View Workload Logs
• View Workload Telemetry
• View Connection Logs
• View Connection Telemetry

Quick Reference¶

Quick Quiz¶

Summary¶

You now have a complete map of the Administration area in the Kelvin UI. Every section has a clear purpose, and they build on each other.

The key things to remember:

• The Dashboard gives you a platform-wide health view with filters — use it regularly
• Asset Management follows a strict hierarchy: Asset Type first, then Assets, then Properties
• Data Streams carry the signals; semantic types and units give those signals meaning
• Data Quality monitors incoming data continuously; the health score is calculated every 5 minutes
• Guardrails are the last line of defence before any value is written to an asset
• Cluster Management is where you manage your edge infrastructure after initial setup
• Connections are deployed workloads that bridge assets to data streams — you manage their IO maps, config, and lifecycle
• User Management controls Kelvin Platform access through Keycloak groups, RBAC, and ABAC
• Telemetry and Logs give you workload and connection health visibility through per-workload Logs and Telemetry tabs, plus a platform-wide Grafana dashboard

For full documentation, visit:

• Administration Dashboard
• Asset Administration
• Data Stream Administration
• Data Quality Dashboard
• Guardrails
• Cluster Management
• Connection Management
• User Management
• Monitor Dashboard
• Grafana Dashboards

Good luck and Happy Learning !