Skip to main content
A cloud run is one execution of an Octoparse task in the cloud. It has its own status, subtasks, data, and run history. A task can have many cloud runs over time; each run must be evaluated separately. The run lifecycle has two related tracks:
  • Execution: whether Octoparse accepted, scheduled, assigned, and ran the workflow.
  • Data readiness: whether extracted records were uploaded, stored, counted, and made available to read or export.
These tracks do not finish at exactly the same time. A run can finish executing while its data is still being stored or its statistics are still updating.

​
How a cloud run progresses

1

A user or schedule starts Cloud Extraction

Start the task from the Task List, or let a cloud schedule start it at the configured time. A batch start can also start several selected tasks.
2

The run enters cloud execution

If cloud capacity is available, the run proceeds to execution. If the available nodes or parallel-task capacity are in use, the run waits in the queue until capacity is released.
3

The workflow is split and assigned

If the task contains work that can run independently, Octoparse creates subtasks and assigns them to available cloud nodes. A task that cannot be split runs as one execution unit.
4

Cloud nodes execute the subtasks

Each assigned node follows the task workflow and reports progress, extracted records, stops, and errors. Subtasks can finish at different times or produce different results.
5

Records are stored as Cloud Data

Extracted records are uploaded and written to Cloud Data. Duplicate handling and data processing can make the saved record count lower than the extracted count.
6

The run becomes ready to read or export

After storage and statistics processing catch up, the records can be read in the Data List, exported, or retrieved through OpenAPI. This step can lag behind the execution status.

​
Run statuses

The Task List and Cloud Run views use these run-level statuses:
A status is an execution signal, not a data-quality guarantee. A Completed run can still have no records, fewer records than expected, duplicates, or a task logic error.

​
Subtask status

Subtasks are the execution units created from a task. Open Subtask Status to see each subtask’s status, start and end time, elapsed time, data extracted, event logs, and screenshots. Use these details to identify which part of a split run is progressing or blocked. The exact subtask labels can vary by run and product view. Interpret them together with the run status and the subtask details rather than treating an internal processing state as a separate user-facing run status. Not every subtask runs at the same time. The account’s cloud nodes, task concurrency, queue priority, and task structure determine how many can be active together. A large number of subtasks therefore does not prove that the run is fully parallel.

​
A completed run may need a moment before data is ready

The run status tells you whether cloud execution has finished. The Data List and export views become available after the extracted records have been processed. For that reason, a run can show Completed before all records appear in Cloud Data. When checking a result, use this order:
  1. Confirm the run shows Completed.
  2. Open Data List and check the record count and fields.
  3. If the list is temporarily empty or incomplete, refresh after a short wait.
  4. If the problem remains, open Subtask Status and review the affected subtask’s details, event log, and screenshot.
Do not treat an empty Data List immediately after completion as proof that the task scraped no data. If the list remains empty after processing time, follow the cloud run monitoring workflow to distinguish a task issue from a data-storage or export issue.

​
Start failures and stopped runs

A run may fail to start, or it may stop after one or more subtasks have begun.

​
If the run does not start

If no cloud run appears after you start the task, or the start action returns an error, check:
  • The task cannot be found or the user does not have permission to run it.
  • Another run of the same task is already active.
  • Cloud Extraction is not enabled for the account.
  • The account or task has insufficient capacity or entitlement.
  • The task configuration or required input is invalid.
Check the message shown by the product and the task or account settings. This is different from a run that appears in the queue and later fails.

​
Stopped after execution begins

Possible causes include:
  • A user stopped the run.
  • The task reached a timeout or stop condition.
  • The target site blocked the request or required CAPTCHA handling.
  • Login, cookies, selectors, pagination, or waits did not work in the cloud.
  • A cloud node or executor failed.
Open the first affected subtask, review its event log and screenshot, and compare its output with a successful run. Restart only after correcting the cause; restarting an unchanged workflow can repeat the same failure.

​
Check a cloud run before scheduling it

For a new or changed task:
  1. Start one cloud run from the Task List.
  2. Check that the run moves from Queued to Running, and that its subtasks make progress.
  3. When the run is Completed, open Data List and compare the fields, record count, and duplicate count with the expected result.
  4. If the result is correct, enable the schedule. If it is not, inspect Subtask Details, logs, and screenshots before running the task again.
For a recurring task, compare each new run with the last known good run. A Completed status confirms that execution ended; it does not by itself confirm that the output is complete or correct.

Cloud resources and queues

Understand node allocation, splitting, queueing, and concurrency.

Monitor cloud runs

Inspect run metrics, data, subtasks, logs, and screenshots.

Automate cloud data exports

Export stored Cloud Data to files, databases, storage, or OpenAPI.

Schedule cloud runs

Configure recurring cloud runs and handle schedule overlap.