Platform Status FAQ

Answers to common questions about the Platform Status page, health indicators, capacity readouts, Dataplane Diagnostics, and restarts.

General

What is the Platform Status page?

A real-time dashboard showing the health and configuration of your Qualytics deployment. It has four cards: Deployment, Components, Dataplane, and, on self-hosted deployments, License. See Platform Status for a full breakdown.

Who can access the Platform Status page?

Users with the Manager or Admin role. Members are redirected away. Restarting the Dataplane and opening Dataplane Diagnostics require the Admin role. See the Permissions page.

What do the status indicators mean?

On the Components card, a green dot means the component responded and is operating normally, a red dot means it did not respond or reported a failure, and a grey dot means the state is unknown because the status request itself did not return.

How often does the status data refresh?

Once, when the page loads. Use Refresh Status Summary to get a current reading.

What happens if the refresh fails?

An error notification appears and the page keeps the values it already had. Try again, or reload the browser page. If it keeps failing, the platform backend may be unreachable.

Components

What happens when the database is unhealthy?

The platform cannot save or read any data. Sign-in, datastore management, and operation history are all affected. Users who are already signed in see the app lock behind a Database Connection Lost dialog until the connection returns; see Platform Outage.

What happens when RabbitMQ is unhealthy?

No operation request can reach the Dataplane. Sync, Profile, Scan, External Scan, Export, and Materialize all become unavailable, and because the Dataplane health check travels the same path, the Dataplane row reports unhealthy too.

Can the database be healthy but RabbitMQ unhealthy?

Yes. They are independent services. The database handles persistence and RabbitMQ handles message routing between the platform and the Dataplane, so one can fail without the other.

Why does a component show a grey dot instead of red?

Grey means unknown, not unhealthy. The status request itself did not come back, so the page has no reading to report. Refresh to try again.

What happened to the Private Routes section?

Reachability checks for datastores on private networks are no longer reported on the Status page. Connectivity to a private host is verified when you test the connection on the datastore itself. See Connection network requirements.

Dataplane

What is the Dataplane?

The service that runs every data quality operation: Sync, Profile, Scan, External Scan, Export, and Materialize. Depending on your deployment it runs on Kubernetes, Databricks, or a local host, and it communicates with the rest of the platform through RabbitMQ.

What does it mean when the Dataplane is unhealthy?

It is not answering health check requests. That can happen during a restart, during a deployment update, or if the engine has crashed. Its configuration fields display -- until it comes back.

What is the Thread Pool State readout?

A live view of the engine's workload, showing how many operations are running and how many are queued, with the running operation IDs listed below as pills. When nothing is running or queued it reads Idle. See Capacity Readouts.

What does a high number of queued requests mean?

All available executor slots are occupied and new operations are waiting. If it stays high, the deployment may need more executors. Contact the Qualytics support team to discuss scaling.

Why do I not see the Executor Capacity or Sync Capacity readouts?

They only appear when the running engine reports them. Deployments on Databricks and engine builds that predate these measurements omit them rather than showing zeros.

The page warns that cluster autoscaling may be misconfigured. What does that mean?

The workload needs more executors than are running, and the engine has not yet reached its configured maximum, so executors are being requested but are not arriving. Check that node pool autoscaling is enabled and that nodes are schedulable. See Capacity Warnings.

How long does a Dataplane restart take?

It depends on your deployment, because the orchestration platform performs the actual restart. The Status page re-checks automatically about 20 seconds after you confirm. Until the engine answers again, the Components card reports it unhealthy and the fields on the Dataplane card's Configuration subtab display --.

What operations are affected during a restart?

Connection tests, schema discovery, dry runs, source record views, and computed asset validation fail immediately. Running operations are interrupted and automatically restarted once the engine is back: Profile, Scan, and Promote continue with the work that had not been processed yet, Sync, Export, and Materialize start over from the beginning, and External Scan is aborted instead. Flows, user management, teams, tags, scheduling, and integrations are unaffected. See Restart Dataplane and the Runs FAQ for details.

What happens to scheduled operations during a restart?

Scheduling configuration stays available, so nothing is lost. An operation that triggers while the engine is restarting is queued and picked up automatically once it is back.

Dataplane Diagnostics

What is a diagnostics report?

A detailed capture of what the Dataplane is doing at that instant: the operations running with their progress, a timeline of each one's steps, memory and task capacity in use, and automated findings that name a slow or unhealthy pattern with its likely cause. See Dataplane Diagnostics.

Who can capture a diagnostics report?

Only Admins. The Diagnostics subtab is not shown to Managers at all.

Why does a capture take so long?

The engine renders the report on demand, while it is still running your workload. A capture can take up to about 30 seconds.

Are diagnostics reports saved?

No. Each capture is live and nothing is retained, so there is no history to browse. Use Download snapshot to keep a copy of the report on screen.

How do I keep evidence of what happened before a restart?

Capture a report and download it before restarting. Restarting clears the state the report describes, and the platform does not keep reports.

Can I capture a report when the Dataplane is already down?

No. A capture requires the engine to answer and render the report. If it has stopped responding entirely, there is nothing to capture.

Should I leave automatic recapture running?

No. Every capture makes the engine render a report while it is working. Use the 5s, 10s, or 30s intervals to watch a live incident, then set it back to Off. It also switches off automatically when you leave the subtab.

What does the "owner" column in the Findings table mean?

Who can act on the finding: source means the data source system, cluster means the compute environment, config means a container setting such as a partition field, and unattributed means the symptom is real but no rule matched the evidence. See Causes and owners.

What does "blast radius" mean?

How far a finding's damage reaches: localized affects only that container, operation drags the whole operation, and cluster means it is holding shared capacity while other work waits. See Blast radius.

Can I share a downloaded report outside my organization?

The report contains container names, source types, versions, and resource figures for your deployment, but no records from your data. Review it before sharing externally, the same as any status summary.

License

Why do I not see the License card?

Because your deployment is managed by Qualytics. The License card renders only on self-hosted deployments. On a managed deployment Qualytics handles licensing for you, so there is nothing to see or do on the Status page. See License.

My deployment is self-hosted and the card shows "No license found". What does that mean?

No license key has been applied yet. The card shows a grace-period countdown instead of an expiration date. A license is required after 31 days of datastore connectivity, and once the countdown runs out, Dataplane operations are restricted until a valid key is applied. See Grace Period.

How do I renew an expiring license?

Use Generate License Request on the License card to produce the request to send to Qualytics, then Update License to apply the key you receive. See Update License.

API

Can I check the platform status via API?

Yes. GET /api/status is public and needs no token, so a monitoring probe can read platform health directly. See the Status API page.

Can I capture diagnostics via API?

Yes. GET /api/dataplane/diagnostics returns the report as HTML and requires the Admin role. See the Status API page.

Can I restart the Dataplane via API?

Yes. PUT /api/admin/restart triggers a graceful restart and requires the Admin role. See the Status API page.