Usage Bundle metrics

The Usage Bundle contains reports about product usage, configuration, settings, and the instance. Use this article to identify the CSV files in a bundle and understand their columns. For instructions to create and send a bundle, see Usage Bundle.

You need the View Usage Reports permission to view or download the bundle.

Files in the bundle

Each bundle includes the standard reports. If you request Product Analytics data, the bundle also includes Product Analytics reports. Automated transmissions can include reports for the previous day. Legacy diagnostic and compatibility exports can include six detailed transcription reports.

Location or file	Contents	Availability
Root folder	Reports at the top level of the bundle.	All bundles
`application_usage`	Daily instance metrics for volume, automation, quality, adoption, and configuration.	All bundles
`environment_configuration`	Environment-variable names and redacted values.	All bundles
`browser_usage_by_day_and_user`	Browser and user-agent data by day and user ID.	All bundles
`field_machine_transcriptions_report` `cell_machine_transcriptions_report` `field_supervision_transcriptions_report` `cell_supervision_transcriptions_report` `checkbox_machine_transcriptions_report` `signature_machine_transcriptions_report`	Detailed transcription records for fields, table cells, checkboxes, and signatures. The reports separate machine and Supervision results.	Legacy diagnostic and compatibility exports only
`product_analytics/` folder	Detailed analytics for product usage, flows, layouts, models, users, and infrastructure.	When Product Analytics data is requested
`infrastructure`	Versions of the database, operating system, Kubernetes, and Helm chart.	Product Analytics
`vlm_gpu_usage`	ORCA Vision Language Model (VLM) request and image counts, plus processing time, by day, task, and base model.	Product Analytics
`image_correction_usage`	Rotation and mobile-capture correction results, model details, request counts, and processing time.	Product Analytics
`settings`	Application settings supplied when the bundle is generated.	Product Analytics; only when settings are supplied
`analytics_metrics` `analytics_metrics_per_day`	Metric-store inventory and rollup data, including namespaces, metric names, counts, dimensions, and time coverage.	Product Analytics
`entry_accuracy` `entry_automation`	Entry accuracy and automation metrics for fields and table cells.	Product Analytics
`classification_automation_metrics` `classification_accuracy` `classification_revisions` `classification_qa_layout_revisions`	Classification automation, accuracy, and layout-revision activity.	Product Analytics
`custom_supervision_chat_usage`	Prompts, responses, and distinct users by day.	Product Analytics
`manual_task_metrics`	Manual task volume and accumulated waiting and active time by day, flow, and layout.	Product Analytics
`throughput_metrics` `hourly_pages_metrics`	Flow- and layout-level throughput by day and completed-page throughput by hour.	Product Analytics
`system_resource_usage` `workers_resource_usage`	CPU and memory use for hosts, plus resource use and lifespan for workers.	Product Analytics
`base_models` `base_model_workflows`	Available base models, installation status, usage counts, and the flows that use each model.	Product Analytics
`block_counts` `connector_counts` `file_filter_block` `file_filter_throughput`	Block and connector counts plus File Filter configuration and throughput.	Product Analytics
`db_entity_counts`	Row counts for the platform's primary database entities.	Product Analytics
`license_keys`	License-key records and their creation dates.	Product Analytics
Flow and configuration CSVs	Flows, latest flow versions, system configuration, free-form configuration, and SDM configuration.	Product Analytics
User, authentication, qualification, and permission CSVs	Users, profiles, OAuth clients, authentication groups, Django groups, qualifications, permissions, and their relationship tables.	Product Analytics
Layout-management CSVs	Field data types, dictionaries, field-customization groups, templates, layout groups, layout variations, tables, tags, releases, potential layouts, and candidate groups.	Product Analytics
ML-model and ground-truth CSVs	Classification, transcription, field-identification, table-identification, text-classification, and Field VLM models and definitions, plus ground-truth inventories.	Product Analytics
`daily_metrics/` folder	Reports for the previous UTC day.	Automated transmissions when daily metrics are requested
`nlc_automation_metrics_prior_day.csv`	Non-structured classification (NLC) Supervision events, including machine and human layouts, confidence, agreement, and event time.	Daily metrics
`reclassified_documents_report_prior_day.csv`	Documents reclassified during the previous UTC day.	Daily metrics

Application usage

application_usage.csv contains one row for each day in the selected period. It reports metrics for the complete instance, not individual flows or layouts. For layout-level metrics, see Layout-level metrics.

Column	Description	Calculation
Date	The reporting day.	Uses the selected date range. For days with no usage, the metric columns contain `0`.
Number of Submissions Created	New Submission IDs created.	One per submission, dated by submission creation. In v43.1 and later, recorded at creation; in v43.0, recorded at completion.
Number of Submissions Completed	Submissions whose status changed to Complete during the period.	Recorded and dated when the submission completes.
Number of Documents Created	New Document IDs created.	One per document. Recorded when the submission completes and dated by submission creation.
Number of Documents Completed	Documents linked to completed submissions.	One per document. Recorded and dated when the submission completes.
Number of Pages Created	Submission pages linked to submissions created during the period.	Dated by submission creation. In v43.1 and later, recorded at creation; in v43.0, recorded at completion.
Number of Pages Completed	Submission pages linked to submissions completed during the period.	Recorded and dated when the submission completes.
Number of Pages Matched to Structured Layout Variations Created	Pages on Structured documents linked to a submission.	One per Structured-document page; dated by submission creation.
Number of Pages Matched to Structured Layout Variations Completed	Pages on Structured documents linked to completed submissions.	One per Structured-document page; dated by submission completion.
Number of Pages Matched to Semi-Structured Layout Variations Created	Pages on Semi-structured documents linked to a submission.	One per Semi-structured-document page. Recorded and dated when the submission completes.
Number of Pages Matched to Semi-Structured Layout Variations Completed	Pages on Semi-structured documents linked to completed submissions.	One per Semi-structured-document page; dated by submission completion.
Number of Pages with Fields on Them Created	Pages containing at least one field or cell on documents linked to a submission.	One per qualifying page. Recorded and dated when the submission completes.
Number of Pages with Fields on Them Completed	Pages containing at least one field or cell on completed or failed documents.	One per qualifying page, dated by submission completion.
Number of Fields Created	Fields on document pages linked to submissions.	Recorded and dated when the submission completes.
Number of Fields Completed	Completed fields on document pages linked to submissions.	Currently populated from the same field count as Number of Fields Created.
Number of Characters Completed	Characters transcribed by a machine or person in completed, non-skipped fields and cells.	Excludes data removed by PII deletion. Each checkbox counts as one character. Dated by submission completion.
Seats	Distinct users who logged in during the day.	Counts distinct user IDs from recorded login activity.
Number of Files Submitted	Files used to create submissions.	Aggregated at submission start and dated by submission creation.
Number of Entries Machine Transcribed	Completed fields and cells transcribed without human involvement.	Recorded for documents linked to submissions and dated by submission completion.
Number of Entries Manually Transcribed	Completed fields and cells whose transcription involved a person.	Recorded for documents linked to submissions and dated by submission completion.
Number of Fields Machine Identified	Completed fields identified by the machine, excluding Flexible Extraction tasks.	Recorded for documents linked to submissions and dated by submission completion.
Number of Fields Manually Identified	Completed fields identified by a person, excluding Flexible Extraction tasks.	Recorded for documents linked to submissions and dated by submission completion.
Number of Pages Classified Automatically	Pages classified by the machine.	Recorded for pages linked to submissions and dated by submission completion.
Number of Pages Classified Manually	Pages classified by a person.	Recorded for pages linked to submissions and dated by submission completion.
Number of Unique Layout Variations Matched	Distinct layouts matched to submission pages.	A submission can contribute on two days when its creation and completion dates differ.
Number of QA Responses on Manual Transcription	Transcription QA responses for human transcriptions without manual Supervision.	Aggregated from QA responses on Structured and Semi-structured documents and dated by submission completion.
Number of QA Correct Responses on Manual Transcription	Correct QA responses for human transcriptions without manual Supervision.	Aggregated from QA responses and dated by submission completion.
Number of QA Responses on System Transcription	System-attributed Transcription QA responses without manual Supervision.	Aggregated from QA responses and dated by submission completion.
Number of QA Correct Responses on System Transcription	Correct system-attributed Transcription QA responses.	Aggregated from QA responses and dated by submission completion.
Number of QA Responses on Machine Transcription	Machine Transcription QA responses without manual Supervision.	Aggregated from QA responses and dated by submission completion.
Number of QA Correct Responses on Machine Transcription	Correct machine Transcription QA responses.	Aggregated from QA responses and dated by submission completion.
Number of QA Responses on Manual Field Identification	Identification QA responses for fields identified by a person on Semi-structured documents.	Aggregated from QA responses and dated by submission completion.
Number of QA Correct Responses on Manual Field Identification	Correct Identification QA responses for fields identified by a person.	Aggregated from QA responses and dated by submission completion.
Number of QA Responses on System Field Identification	System-attributed Identification QA responses on Semi-structured documents.	Aggregated from QA responses and dated by submission completion.
Number of QA Correct Responses on System Field Identification	Correct system-attributed Identification QA responses.	Aggregated from QA responses and dated by submission completion.
Number of QA Responses on Machine Field Identification	Identification QA responses for fields identified by the machine.	Aggregated from QA responses and dated by submission completion.
Number of QA Correct Responses on Machine Field Identification	Correct Identification QA responses for fields identified by the machine.	Aggregated from QA responses and dated by submission completion.
Number of Incremental Fields Machine Transcribed	Machine field transcriptions automated through fine-tuning that would otherwise require Supervision.	Recorded and dated when the submission completes.
Organize Documents: Number of Tasks Performed	Completed Classification Supervision tasks.	One per completed task, dated by task completion.
Organize Documents: Number of Pages Shown in Step 1	Unmatched pages sent to completed Classification Supervision tasks.	Aggregated from the task response and dated by task completion.
Organize Documents: Number of Pages Categorized in Step 1	Previously unmatched pages that a person classified.	Aggregated from the task response and dated by task completion.
Organize Documents: Number of Documents Created in Step 1	Documents a user created or changed during Classification Supervision.	Aggregated from the task response and dated by task completion.
Organize Documents: Number of Documents Outputted in Step 1	Documents resulting from completed Classification Supervision tasks.	Aggregated from the task response and dated by task completion.
Organize Documents: Number of Folders Created in Step 2	Folders resulting from completed Classification Supervision tasks.	Aggregated from the task response and dated by task completion.
Number of Live Layout Variations	Live layout variations recorded for the day.	Uses the daily recorded gauge value.
Software Version	The software version running that day.	Uses the version recorded for the reporting interval.
Number of Machine Match Non-Structured Pages	Non-Structured pages classified by the machine.	Recorded for pages linked to submissions and dated by submission completion.
Number of Human Match Non-Structured Pages	Non-Structured pages classified by a person.	Recorded for pages linked to submissions and dated by submission completion.
Number of Releases	Releases recorded for the day.	Uses the daily recorded gauge value.
Number of Archived Releases	Archived releases recorded for the day.	Uses the daily recorded gauge value.
Number of Layout Variations	Layout variations recorded for the day.	Uses the daily recorded gauge value.
Number of Archived Layouts	Archived layouts recorded for the day.	Uses the daily recorded gauge value.
Number of Layout Versions	Layout versions recorded for the day.	Uses the daily recorded gauge value.
Number of Live IDP Flows	Distinct active releases deployed in live IDP flows.	Uses the daily recorded gauge value.
Number of Table Cells Created	Table cells on document pages linked to submissions.	Recorded and dated when the submission completes.
Number of Table Cells Completed	Completed table cells on document pages linked to submissions.	Currently populated from the same cell count as Number of Table Cells Created.
Number of Table Cells Machine Identified	Completed cells identified by the machine, excluding Flexible Extraction tasks.	Recorded for documents linked to submissions and dated by submission completion.
Number of Table Cells Manually Identified	Completed cells identified by a person, excluding Flexible Extraction tasks.	Recorded for documents linked to submissions and dated by submission completion.
Number of QA Responses on Manual Non-Structured Classification Pages	Classification QA responses performed by a person for non-Structured documents.	Aggregated when Classification QA consensus is reached and dated by submission completion.
Number of QA Correct Responses on Manual Non-Structured Classification Pages	Correct Classification QA responses performed by a person.	Aggregated when Classification QA consensus is reached and dated by submission completion.
Number of QA Responses on System Non-Structured Classification Pages	System-attributed Classification QA responses for non-Structured documents.	Aggregated when Classification QA consensus is reached and dated by submission completion.
Number of QA Correct Responses on System Non-Structured Classification Pages	Correct system-attributed Classification QA responses.	Aggregated when Classification QA consensus is reached and dated by submission completion.
Number of QA Responses on Machine Non-Structured Classification Pages	Classification QA responses performed by the machine.	Aggregated when Classification QA consensus is reached and dated by submission completion.
Number of QA Correct Responses on Machine Non-Structured Classification Pages	Correct Classification QA responses performed by the machine.	Aggregated when Classification QA consensus is reached and dated by submission completion.
Number of Fields Extracted in Flexible Extraction	Completed fields transcribed through Flexible Extraction.	Recorded for documents linked to submissions and dated by submission completion.
Number of Table Cells Extracted in Flexible Extraction	Completed table cells transcribed through Flexible Extraction.	Recorded for documents linked to submissions and dated by submission completion.
Number of Fields Extracted in Custom Supervision	Completed fields transcribed through Custom Supervision.	Recorded for documents linked to submissions and dated by submission completion.
Number of Table Cells Extracted in Custom Supervision	Completed table cells transcribed through Custom Supervision.	Recorded for documents linked to submissions and dated by submission completion.
Number of Flows	Flows recorded for the day.	Daily average of the recorded flow count, truncated to a whole number.
Number of Live Flows	Live flows recorded for the day.	Daily average of the recorded live-flow count, truncated to a whole number.
Number of QA Responses on Machine Full Page Transcription	Machine Full Page Transcription QA responses.	Aggregated when QA consensus is reached and dated by submission completion.
Number of QA Correct Responses on Machine Full Page Transcription	Correct machine Full Page Transcription QA responses.	Aggregated when QA consensus is reached and dated by submission completion.
Checksum	An MD5 hash used to detect changes to the row.	Calculated from the row's date and metric values after all other columns are populated.

Layout-level metrics

Use these Product Analytics reports to measure performance by layout. Unless otherwise noted, each row groups data by reporting period, flow, and layout. Use Layout UUID as the stable identifier, Layout Name for display, and Layout Type to distinguish layout types.

`classification_automation_metrics.csv`

Reports classification volume and automation by flow and layout.

Column	Description
Period Start	Start of the reporting period.
Flow UUID / Flow Name	Flow associated with the activity.
Layout UUID / Layout Name / Layout Type	Layout dimensions for the reported pages.
Pages Count	Total pages represented by the row.
Machine Classified Pages	Pages classified by the machine.
Human Classified Pages	Pages classified by a person.
Supervision Pages	Pages handled through Classification Supervision.
QA Pages	Pages handled through Classification QA.

`entry_accuracy.csv`

Reports QA totals and correct-result counts. To calculate accuracy, divide a Correct value by its corresponding Total value. Do not calculate accuracy when Total is 0.

Columns	Description
Period Start	Start of the reporting period.
Flow UUID / Flow Name	Flow associated with the activity.
Layout UUID / Layout Name / Layout Type	Layout dimensions for the activity.
Field Identifications Machine QA - Total / Correct	Machine field-identification QA responses and correct responses.
Field Identifications Human QA - Total / Correct	Human field-identification QA responses and correct responses.
Table Cell Identifications Machine QA - Total / Correct	Machine table-cell-identification QA responses and correct responses.
Table Cell Identifications Human QA - Total / Correct	Human table-cell-identification QA responses and correct responses.
Field Transcriptions Machine QA - Total / Correct	Machine field-transcription QA responses and correct responses.
Field Transcriptions Human QA - Total / Correct	Human field-transcription QA responses and correct responses.
Table Cell Transcriptions Machine QA - Total / Correct	Machine table-cell-transcription QA responses and correct responses.
Table Cell Transcriptions Human QA - Total / Correct	Human table-cell-transcription QA responses and correct responses.

`manual_task_metrics.csv`

Reports manual-task demand and effort. Each task type includes Tasks Created, Tasks Completed, Waiting, and Active columns. Waiting and Active contain accumulated durations.

Columns	Description
Period Start	Start of the reporting period.
Flow UUID / Flow Name	Flow associated with the tasks.
Layout UUID / Layout Name / Layout Type	Layout dimensions for the tasks.
Document Classification - Tasks Created / Completed / Waiting / Active	Document Classification task volume and accumulated time.
Field Identification - Tasks Created / Completed / Waiting / Active	Field Identification task volume and accumulated time.
Table Identification - Tasks Created / Completed / Waiting / Active	Table Identification task volume and accumulated time.
Field Transcription - Tasks Created / Completed / Waiting / Active	Field Transcription task volume and accumulated time.
Table Transcription - Tasks Created / Completed / Waiting / Active	Table Transcription task volume and accumulated time.
Flexible Extraction - Tasks Created / Completed / Waiting / Active	Flexible Extraction task volume and accumulated time.
Custom Supervision - Tasks Created / Completed / Waiting / Active	Custom Supervision task volume and accumulated time.

`throughput_metrics.csv`

Reports completion time and processed volume by flow and layout.

Column	Description
Period Start	Start of the reporting period.
Flow UUID / Flow Name	Flow associated with the processed work.
Layout UUID / Layout Name / Layout Type	Layout dimensions for the processed work.
Submission Time to Completion	Submission processing time to completion.
Document Time to Completion	Document processing time to completion.
Submissions	Number of submissions represented by the row.
Documents	Number of documents represented by the row.
Pages Processed	Number of processed pages.
Pages Submitted	Number of submitted pages.
Fields	Number of fields represented by the row.
Table Cells	Number of table cells represented by the row.

`classification_accuracy.csv`

Reports flow-level classification QA totals and correct-result counts. This report does not contain Layout UUID, Layout Name, or Layout Type columns.

Column	Description
Period Start	Start of the reporting period.
Flow UUID / Flow Name	Flow associated with the activity.
Classification Machine QA - Total / Correct	Machine-classification QA responses and correct responses.
Classification Human QA - Total / Correct	Human-classification QA responses and correct responses.
System Classification Total / Correct	System-attributed classification QA responses and correct responses.

`classification_revisions.csv`

Records human changes to machine-selected layouts.

Column	Description
Period Start	Start of the reporting period.
Flow UUID / Flow Name	Flow associated with the classification.
User ID	User associated with the revision.
Machine Layout UUID / Machine Layout Name	Layout selected by the machine.
Revised Layout UUID / Revised Layout Name	Layout selected during human revision.
Page Count	Number of pages represented by the row.
Average Machine Confidence	Average confidence of the machine classifications represented by the row.

`classification_qa_layout_revisions.csv`

Records differences between the layout selected by non-structured classification (NLC) and the Classification QA consensus.

Column	Description
Period Start	Start of the reporting period.
Flow UUID / Flow Name	Flow associated with the classification.
NLC Layout UUID / NLC Layout Name	Layout selected by non-structured classification.
QA Consensus Layout UUID / QA Consensus Layout Name	Layout selected by QA consensus.
Page Count	Number of pages represented by the row.

VLM GPU usage

In v43.2 and later, product_analytics/vlm_gpu_usage.csv reports ORCA request and image counts, plus processing time. Each row groups data by day, task, and base model.

Column	Description
Period Start	The reporting date.
Task Name	The ORCA task, such as `field_extraction` or `general_prompting`.
Base Model ID	The base model used for the task.
Processing Seconds Avg	Average processing time for the reported requests.
Processing Seconds Max	Longest processing time for the reported requests.
# of requests	Number of ORCA requests.
# of images	For Field Extraction, the number of document pages processed; for general prompting, the number of images submitted.

Infrastructure

product_analytics/infrastructure.csv contains one row for each platform component in the report.

Component	Reported version
`database`	Database vendor and version.
`system`	Operating-system platform and version.
`kubernetes`	Deployed Kubernetes version.
`helm-chart`	Helm chart version used for deployment.