Skip to content

Metrics

UbiOps generates metrics based on activity in the platform.

The recommended way to access the metrics is through the WebApp (available at https://app.ubiops.com for our SaaS solution). It is also possible to access the logs via CLI or Client libraries, see Interfacing with UbiOps.

Metrics can be found under Monitoring in the left navigation panel. This will show an overview of project level metrics. It is possible to filter metrics on date, or on properties like deployment_name.

Many objects in UbiOps (like deployments and pipelines) have a Metrics button which will show the metrics relevant to the object.

The following metrics are available.

Deployment metrics

Metric1 Unit Description
active_time seconds (float) Time in seconds that the deployment is active
gb_seconds seconds (float) Usage of GB seconds 2
instances instances (float) Average number of active deployment instances
input_volume bytes (int) Volume of incoming data in bytes
output_volume bytes (int) Volume of outgoing data in bytes
compute3 seconds (float) Average time in seconds for a request to complete
memory_peak bytes (int) Peak memory used during a request
requests requests (int) Number of requests made to the object
failed_requests requests (int) Number of failed requests made to the object
outputs items (int) Number of outgoing data items

Pipeline metrics

Metric1 Unit Description
requests requests (int) Number of requests made to the object
failed_requests requests (int) Number of failed requests made to the object
request_duration3 seconds (float) Average time in seconds for a pipeline request to complete
input_volume bytes (int) Volume of incoming data in bytes
output_volume bytes (int) Volume of outgoing data in bytes
object_requests requests (int) Number of requests made to deployments in a pipeline
object_failed_requests requests (int) Number of failed requests made to deployments in a pipeline

  1. Deployment and pipeline metrics are sampled every 60s 

  2. gb_seconds is calculated by multiplying the deployment memory sizes in GB by the active_time and instances 

  3. compute is the actual time spend on calling the 'request' function of a deployment. In contrast request_duration is the actual time spent on executing a pipeline request and also depends on the shape of a pipeline (e.g. parallel or serial deployments)