# Monitor and operate CDC

Use the SAP HANA RFC Endpoint Cockpit to monitor registered tables, capture jobs, and task subscriptions.

## Monitor subscriptions

In the SAP HANA RFC Endpoint Cockpit, double-click a table or click Details. The table detail view has Subscriptions, Jobs overview, Change Capture Index, and Diagnostics tabs.

Subscription fields

| Field | Description |
| --- | --- |
| `SUBSCRIBER_ID` | Unique identifier for the task's subscription to the table. |
| `INIT_STREAM_TS` | Timestamp when the subscription started streaming. |
| `LAST_CONSUMED_TS` | Last point through which the task consumed and acknowledged changes. |
| `PACKAGE_SI` | Package size configured for the subscription. |
| `CDC_STATUS` | `ACTIVE` while the table is being polled; `PAUSED` when no task is currently polling it. |

Information note

All subscribers to a table share one changelog and one cleanup boundary. SAP removes a change only after every subscriber has acknowledged it and the retention period has elapsed. A stopped or lagging task can prevent cleanup for every other task that reads the table. Check whether `LAST_CONSUMED_TS` is advancing.

## Restart or resynchronize CDC

To resume capture from a specific point, stop the task and choose Start from timestamp, or start a reload. Do not use a plain resume when you need to establish a new starting point. Starting from a timestamp creates a new SAP subscription for that table and replaces its previous subscription.

## Remove subscriptions

* Remove a table from a task: The subscription is deleted the next time you resume or reload the task. Other tables in the task are not affected.
* Delete a task: Its subscriptions remain in SAP. Remove them manually in the SAP admin transaction.

Information note

A leftover subscription stops advancing its consumed timestamp and can prevent cleanup for other tasks using the same table. Remove subscriptions for deleted tasks.

## Troubleshoot failures

Common endpoint issues

| Symptom | What happens | Action |
| --- | --- | --- |
| Task stops with an authorization error | The error is fatal and is not retried. | Grant the SAP user the required add-on authorization. See [Required permissions](https://help.qlik.com/en-US/replicate/May2026/Content/Replicate/Main/SAP%20HANA%20RFC/SAP_HANA_RFC_permissions.htm). |
| Subscription fails for some tables | The user is not authorized for those tables. | Grant the required authorization for each table. |
| A table is suspended as not active | SAP reports that the table is not active. The task continues running. | The recovery action for this response is not specified. Contact your Qlik Early Access contact. |
| A table is suspended with a reload-required message | SAP sent a resynchronization marker. Changes are acknowledged up to the marker, then the table is suspended. | Reload the table. |
| Change discovery is postponed after a timeout or dropped connection | The table remains unresolved and is retried in the next cycle. Reads, acknowledgements, and checkpointing are skipped while the lookup is outstanding. | Check whether the table recovers. If it does not, investigate the connection and SAP-side job logs. |

### Transient RFC failures

For an RFC timeout or dropped connection during polling, Qlik Replicate retries the call four times, with waits of 1 to 8 seconds between attempts. If the call still fails, the error is reported. Failures while decoding or applying changes locally are not retried.