Skip to content

DeltaFi Metrics

A DeltaFi deployment provides normalized metrics generation and centralized metrics collection and visualization. Metrics are generated and sent to a VictoriaMetrics time-series database that is used as a data source for Grafana visualization.

Default Metrics Generation

DeltaFi automatically generates metrics for action cognizance, ingress and egress volume, and system utilization. These metrics can be queried from VictoriaMetrics or referenced by Grafana when creating metrics visualizations.

Metrics Lexicon

The following metrics are generated to VictoriaMetrics by default:

NameTagsDescription
stats_counts.bytes_inaction, ingressFlow, sourceNumber of bytes ingressed by a flow (only produced on action='ingress')
stats_counts.files_inaction, ingressFlow, sourceNumber of files ingressed by an action (action='ingress' provides the ingressed files)
stats_counts.files_erroredaction, ingressFlow, sourceNumber of files errored (action='ingress' provides the ingressed files)
stats_counts.files_filteredaction, ingressFlow, sourceNumber of files filtered (action='ingress' provides the ingressed files)
stats_counts.files_droppedaction, ingressFlow, sourceNumber of files dropped (action='ingress' provides the ingressed files)
stats_counts.bytes_outaction, ingressFlow, source, dataSink, destinationNumber of bytes egressed (only produced on action='egress')
stats_counts.files_outaction, ingressFlow, source, dataSink, destinationNumber of files egressed (only produced on action='egress')
gauge.action_queue.queue_sizequeue_nameAn instantaneous gauge for the size of an action_queue (number of DeltaFiles waiting on the queue)
gauge.node.disk.limithostname, serviceAn instantaneous gauge for the maximum aggregate storage limit per node and service partition (bytes)
gauge.node.disk.usagehostname, serviceAn instantaneous gauge for current storage utilization per node and service partition (bytes)
stats_counts.survey.bytessurveyFlow, surveyDirectionNumber of bytes surveyed on a flow
stats_counts.survey.filessurveyFlow, surveyDirectionNumber of files surveyed on a flow
stats_counts.survey.subflow.bytessurveyFlow, surveySubflow, surveyDirectionNumber of bytes surveyed on a subflow
stats_counts.survey.subflow.filessurveyFlow, surveySubflow, surveyDirectionNumber of files surveyed on a subflow

Custom Metrics in Action Kits

Each action kit has a custom metric interface built in to the Result class for the action. To add a new custom metric from the load action, simply use the add interface to create custom metrics with custom tags. Note that the actual metric name in VictoriaMetrics will be prepended with stats_counts. and custom metrics will automatically be tagged with action, ingressFlow, and source.

For example, in the Java Action Kit:

java
result.add(
    new Metric("my.custom.metric, 42)
        .addTag("my.tag.1", "important-thing")
        .addTag("my.tag.2", "interesting-thing")
);

Use caution when generating custom metrics. Each custom metric name-tags permutation will take up space in the time series database.

Metrics Visualization

Grafana is used to generate rich visualizations for the metrics contained in VictoriaMetrics. The following default dashboards are included in a DeltaFi installation:

  • System Overview - A general system status dashboard that includes ingress and egress metrics, files errored/dropped/filtered, queue status, action activity, system alerts, and disk utilization trends
  • Flow Summary - A view of the ingress and egress activity for one or more flows
  • Last Seen on Flows - A list of flows with the most recent flow activity
  • Processing Report - A flow level report on flow activity for a given time period

In addition, there are Dashboards with logs, audit logging information, and logging per action.

Metrics APIs

DeltaFi exposes the full API for VictoriaMetrics and Grafana, allowing for easy custom querying, scripting, and metrics injection.

In addition, this API provides summary node info:

Contact US