Skip to content

Infrastructure Discovery

This workflow describes how infrastructure records get into Clarity Transform through its integrations. You store a credential, add an endpoint on an integration's page, let the endpoint's scheduled job collect data, and then review what was discovered in the CMDB and the Data Analysis dashboards. Each stage below covers only what the workflow needs; follow the links for the full screen descriptions.

Stage 1: store a credential

Settings → Identity & Access Management → Credentials

  1. Select the add icon on the grid toolbar. The Create new entry form opens.
  2. Complete Name, User Name, Password and Description. Password is a masked field and Description is a multi-line text box.
  3. Select Save.

A credential has only these fields. The endpoint form in stage 3 picks a credential from this list, so create it first for any integration whose endpoint form asks for a Credential. RVTools and Nutanix Collector endpoints are file uploads and do not use a credential.

If you later edit the credential and change its Name or User Name, re-enter the Password too, or the form does not save.

See Credentials.

Stage 2: choose the integration

Discover → Integrations

The page header is Technology Integrations. The page has the Cards and Table views and two search boxes, Search by Name and Search by Category. Find the integration you need and select its card. The card opens that integration's page, which has two tabs: Endpoints and History (Nutanix SSP also has Operations).

CSV Import, CSV Export and NetFlow open their own pages instead of an endpoint page. For loading records from a CSV file, see CSV Import. NetFlow capture files are uploaded and then imported with Run Import on the NetFlow page: see NetFlow.

See Integrations.

Stage 3: add an endpoint

On the integration's Endpoints tab:

  1. Select the add icon (tooltip Create New Integration Endpoints) on the grid toolbar. The Create new entry form opens. No field is marked as required.
  2. Complete the form. On most integrations it has Name, End Point, Credential, Job Schedule and Enabled:
    • End Point: the help text reads "FQDN or IP Address of management system for technology e.g. vCenter, Nutanix Prism".
    • Credential: a list of the credentials stored in stage 1. The help text reads "Credentials can be configured under Settings/Identity and Access Management/Credentials."
    • Job Schedule: the help text reads "Cron schedule – Use five fields: 〈mm〉 〈hh〉 〈day-of-month〉 〈month〉 〈weekday〉". For example, 00 22 * * * runs at 22:00 every day (illustrative).
    • Enabled: a checkbox, unchecked when the form opens.
  3. Select Save.

You do not choose the technology: the integration page sets it from the card you selected.

Some integrations use a different form:

  • Windows, Linux and Evolve: Name, Credential, Job Schedule and Enabled, with no End Point.
  • SQL Server: Name, Discovery Method, Credential, Job Schedule and Enabled. End Point appears only when Discovery Method is ipRangeScan or manualTargets. Discovery Method offers inventorySignals (the default), browserEnumeration, ipRangeScan and manualTargets. Its help text reads "Select how SQL Server instances should be discovered for this endpoint configuration."
  • RVTools and Nutanix Collector: Name, File Upload (a Choose file... button) and Enabled, with no End Point, Credential or Job Schedule. When you choose the file, the form shows "File uploaded:" and its name. These endpoints have no schedule: run them with Mark as Running in stage 4.

After Save, the toast "Changes saved successfully!" confirms the endpoint was stored. "Error saving changes:" gives the reason it was not, and the form stays open.

Stage 4: run and monitor discovery

Each endpoint row on the Endpoints tab shows Last Run and Last Result, and has the row actions View Details, Mark as Running, Cancel Job and Show Logs. There is no connection-test action.

  1. To run an endpoint now rather than wait for its Job Schedule, select Mark as Running on its row. It is unavailable while Enabled is cleared, while a run is in progress, and while Last Updated is empty, so tick Enabled and save the endpoint first.
  2. Read Last Run and Last Result when the run finishes. Select Reload table data on the grid toolbar to fetch the latest values. Cancel Job cancels a run.
  3. Where Last Result reports errors or warnings, select its info button (View errors or View warnings). The Endpoint Issues panel lists the errors and warnings recorded for the latest run. While a run is still in progress and not yet recorded, it shows only what your browser tab saw live, and says so.
  4. For the full log, select Show Logs. It opens Logging → Event Log in a new browser tab, not filtered to the endpoint, so find the endpoint's entries yourself with the Event Log's search and filters.

The History tab lists past runs with the columns Name, Success Count, Warning Count, Error Count, Time Taken, Date and Result.

How to read Last Result on the Endpoints tab:

Last Result What it tells you
Empty The endpoint has not run yet.
Starts with Completed, Success, Failed or Cancelled The last run has finished. Mark as Running is available again (if Enabled is ticked).
Anything else Clarity treats a run as still in progress, so Mark as Running stays greyed out.
Contains "error" or "failed" An info button, View errors, opens the Endpoint Issues panel.
Contains "warning" (and no error) An info button, View warnings, opens the same panel.

On the History tab, Result values seen in this build are Success, Completed with warnings and Completed with errors.

Scheduled jobs are listed at Settings → System → Background Jobs. Most of them are integration discovery jobs named after their endpoints. The grid shows each job's Schedule (a five-field cron expression), Last Run, Next Run and Last Result, for example Completed at <timestamp> - Success or Failed at <timestamp> - <error text>. This grid has no run-now action. In this build, Status showed Pending for every job.

See Background Jobs and Event Log.

Stage 5: review the records in the CMDB

Discover → CMDB

Select a CI type card, for example Device, Host, Cluster or Datacenter, to load its grid on the same page. In each grid, Integration names the integration that supplied the record. The row action View Details opens a read-only side panel with the record's populated fields. On these four grids the download menu offers Export Visible Fields and Export All Fields.

The grid remembers its Search: text, column filters, sort order and page size between visits. If the grid shows fewer rows than you expect, clear the Search: box and the filter row.

See CMDB.

Stage 6: review the results in Data Analysis

Discover → Data Analysis → Infrastructure Discovery opens the Infrastructure Discovery Dashboard, with the tabs Executive Summary, Compute Infrastructure, Devices, Databases and Rightsizing. See Infrastructure Discovery.

Discover → Data Analysis → Discovery Sources shows "Per-source discovery health and what each integration contributed to the CMDB". Select a source in the Sources panel on the left. The right-hand panel shows a Run health summary above two tabs, What was discovered and Hierarchy. See Discovery Sources.

Before you trust the estate figures, check STALE RECORDS and Feeds for each source on Discover → Data Analysis → Discovery Sources. A high STALE RECORDS count means the CMDB holds records the platform no longer returns. See Discovery Sources.

Troubleshooting

Symptom Cause What to do
The Credential list on the endpoint form is empty, or lacks the account you need The credential has not been created yet. Add it under Settings → Identity & Access Management → Credentials (stage 1), then open the endpoint form again.
Editing a credential will not save: "Please re-enter the password if you changed the credential name." or "...if you changed the user name." Name or User Name was changed and Password was left as it was. Type the password again in Password and select Save.
The endpoint form has no End Point field The integration does not take one: Windows, Linux, Evolve, RVTools and Nutanix Collector, and SQL Server unless Discovery Method is ipRangeScan or manualTargets. Nothing to fix. For SQL Server, choose ipRangeScan or manualTargets if you need to give targets.
"Error saving changes:" after Save on the endpoint form The save was rejected. The text after the colon gives the reason. Correct the field it names and select Save again. The form stays open, so nothing is lost.
Mark as Running is greyed out Enabled is cleared, Last Updated is empty, or Last Result shows a run still in progress. Tick Enabled and save the endpoint. If a run is in progress, wait for it or select Cancel Job, then Reload table data.
Mark as Running greys out and then becomes available again, with no message The request to start the run failed. Select Reload table data, check Last Result, and look for the reason on the Event Log.
Last Run and Last Result have not changed after a run The grid has not been reloaded since the run finished. Select Reload table data on the grid toolbar.
Last Result reports errors The run hit errors. Select View errors to open the Endpoint Issues panel, then the History tab for the counts, then Show Logs for the detail.
The Endpoint Issues panel says "The latest run has not been recorded yet. Showing only what this browser tab saw live." The run is still in progress, or its record has not been written. Wait for the run to finish, select Reload table data, and open the panel again.
The Endpoint Issues panel says "No error or warning details were recorded for the latest run." The latest run recorded no itemised errors or warnings. Check Error Count on the History tab, and the run's entries on the Event Log.
Show Logs opens a long, unfiltered log Show Logs opens the Event Log without filters. Set Date to the day of the run, then narrow with Level, Quick Time Range and Search:.
The run succeeded but the CMDB has no new records The run collected nothing, often because the account cannot see the objects. Check the account's access against the integration's own page, then run the endpoint again.
The grid shows fewer endpoints, jobs or log entries than expected The grid remembers its Search: text and column filters between visits. Clear the Search: box and the filter row.

Questions people ask

Does saving an endpoint start discovery? Saving stores the endpoint. A run starts when the Job Schedule falls due or when you select Mark as Running. Do not wait for the schedule: run every new endpoint once by hand.

Is there a "Test connection" button? No. The first Mark as Running is the test. Read Last Result and View errors afterwards.

Can I pick Daily or Weekly instead of writing cron? No. Job Schedule takes a five-field cron expression only, for example 0 2 * * * for 02:00 every day.

How do I stop an endpoint running? Cancel Job stops the current run. To stop it running again, clear Enabled on its row, or remove the endpoint. Removing it also removes its background job and that job's history.

Where do I see when it will next run? On Settings → System → Background Jobs, in Next Run for the job named after the endpoint.

Does Show Logs take me to this endpoint's log entries? No. It opens the whole Event Log in a new tab. Filter it yourself.

Which source wins when two integrations report the same field? The one with the higher Integration Priority, set in the Table view of Discover → Integrations. A change applies to future runs, not to records already written.