Platform Tools
Logo
Open Telemetry

Open Telemetry

Logo

3 mins READ

Overview

UnifyApps' OpenTelemetry feature enables users to publish telemetry metrics to Prometheus, allowing for enhanced analytics and logging of platform features. These metrics provide valuable insights into the performance and behavior of key platform components, supporting better monitoring, debugging, and decision-making processes.

Frame 119 (2).png
Frame 119 (2).png

Telemetry metrics can be published across the following three asset classes:

  1. Automations: Metrics can be configured at various levels, including:

    • Node Level: Metrics for individual node performance.

    • Individual Automation Level: Metrics for specific automations.

    • Global Automation Level: Metrics for all automations across the platform.

  2. Connections: Metrics can be published at:

    • Individual Connection Level: Metrics for specific connections.

    • Global Connections Level: Metrics for all connections on the platform.

  3. APIs: Metrics can be configured at:

    • API Group Level: Metrics for specific API groups.

    • Global APIs Level: Metrics for all APIs on the platform.

Prometheus collects the published metrics for your UnifyApps environment. You don't connect a Prometheus account in the product. To query the metrics, use the Prometheus that monitors your environment, and ask your platform administrator for access if you don't have it.

Key Features of Telemetry in UnifyApps

  1. Metric (Counter or Timer):

    • Metrics can be configured in two formats:

      • Counter: Tracks the number of occurrences for a specific action or event.

      • Timer: Measures the time taken for a specific action or event. Optionally, click Add Bucket to add bucket values (non-negative numbers or decimals, in milliseconds). The timer is then published as a histogram.

        Frame 118
        Frame 118

  2. Publish If:

    • Metrics can be configured to publish for different events across the asset classes

    • For Automations:

      • Always on Finished Execution: Metrics are published after the completion of an execution, regardless of its outcome.

      • Always at Start of Execution: Metrics are published at the beginning of an execution. This option is available for Counter metrics only, because a duration can't be measured before the run finishes.

      • Only on Successful Execution: Metrics are published only when the execution is completed successfully.

      • Only on Failed Execution: Metrics are published when the execution fails or is cancelled.

    • For Connections:

      • Always: Metrics are published for every connection event.

      • Success: Metrics are published when a connection event succeeds.

      • Error: Metrics are published when a connection event encounters an error.

    • For APIs (the dropdown shows these as ALWAYS, SUCCESS and FAILURE):

      • Always: Metrics are published for every API event.

      • Success: Metrics are published when the API responds with a status below 400.

      • Failure: For APIs, metrics are published when the API responds with a status of 400 or above, or doesn't respond.

    • Frame 117
      Frame 117

  3. Tags:

    • Users can configure key-value pairs corresponding to the asset class, allowing customization of logged data.

    • For each tag:

      • Tags can be added as keys, and values are logged for each event.

      • Datapills

        can be mapped into a tag's value. Enter the key as plain text, then type {{ in the Value to pick a datapill. Each event then logs a dynamic value for that tag. This enables differentiation between events and provides detailed insights into the event logs.

        Frame 116
        Frame 116

Available tags-values corresponding to the asset classes:

Asset Class

Level

Fields Available

1. Automations

Node Level

App Name, Tenant ID, Workflow Title, Resource Name, User ID, Node ID, Workflow ID, Execution Status, Deployed Version, Application Id, Packaged Application Id

Automation Level

Tenant ID, Workflow Title, User ID, Workflow ID, Execution Status, Deployed Version, Application Id, Packaged Application Id

Global Level

Tenant ID, Workflow Title, User ID, Workflow ID, Execution Status, Deployed Version, Application Id, Packaged Application Id

2. Connections

Connection Level

App Name, Connection Id, Resource Name, Connection Name, Outcome, Response Status, Tenant Id, User Id, Application Id, Packaged Application Id, Workflow Id, Workflow Title

Global Level

App Name, Connection Id, Resource Name, Connection Name, Outcome, Response Status, Tenant Id, User Id, Application Id, Packaged Application Id, Workflow Id, Workflow Title

3. APIs

API Group

Collection Path, Endpoint ID, Http Method, User ID, API Path, Http Status, Tenant ID, Host, Cache Hit, API Client ID, Auth Type, Collection ID, Outcome, Endpoint Name, Collection Name

Global Level

Collection Path, Endpoint ID, Http Method, User ID, API Path, Http Status, Tenant ID, Host, Cache Hit, API Client ID, Auth Type, Collection ID, Outcome, Endpoint Name, Collection Name

Where to configure telemetry

Level

Where to configure it

Global automation level

Automations > Settings > Telemetry. To edit it, you need permission to view settings and to create automations. Otherwise it's read-only.

Individual automation level

The automation's Settings > Telemetry Metrics. Turn on Enable telemetry for this automation.

Node level

The step's settings. Select Enable Telemetry.

Global connections level

Connections > Settings

Individual connection level

The connection's Settings tab

Global APIs level

API Manager > Settings > Open Telemetry Metrics

API group level

The API group's Settings > Open Telemetry Metrics

For APIs, the Metric list offers three metrics: Api Execution - Counter, Api Execution Duration - Timer and Api Resource Execution Duration - Timer. For what each API metric measures, see Open Telemetry Metrics.

Add a metric

  1. Open the telemetry settings for the level you want (see the table above). For an automation or a step, turn on telemetry first.

  2. Click Add New Metric (Add Metric on a step).

  3. In Metric Name, enter a name using letters, digits and underscores only. The name can't start with a digit (for example, api_calls). The name, Metric and Publish If are all required.

  4. In Metric, select a counter or a timer. In an automation's own Settings tab, this field is labelled Aggregation.

  5. In Publish If, select when the metric is published. In an automation's own Settings tab, this field is labelled When to Capture.

  6. Optionally, under Tags, add key-value pairs. Type {{ in a Value to map a datapill.

  7. For a timer, optionally click Add Bucket to add bucket values.

  8. Save your changes.

How metrics reach Prometheus

Automations, connections and APIs send each metric to a shared metrics service. Prometheus polls that service at a regular interval, so a new value can take a short while to appear.

Each tag you add becomes a Prometheus label, and every metric also carries a customerId label. For example, a counter named api_calls appears in Prometheus as api_calls_total.

Frame 115.png
Frame 115.png