Network Discovery scans help you paint a clear picture of your environment. When they run smoothly, they feel like a steady pulse across your network. But when a manual scan suddenly interrupts a scheduled scan, the rhythm breaks. This post explains why this happens and how you can keep your discovery cycle flowing without surprises.
What is going on
Some Lansweeper Cloud users report that starting a manual scan while a large scheduled Discovery Action is running causes the current queue to clear. The scheduled scan appears to stop, and scanning does not resume until the next scheduled trigger. This creates confusion, especially during large scans that take hours to complete.
The root cause lies in how the Discovery Hub and Sensor engine manage scan actions. The scan queue and progress tracking live only in memory. When a new scan is triggered, the Sensor evaluates whether the new action contains more detailed scan options. If it does, it cancels the current scan so it can run the action expected to produce richer results. This is intentional.
For example, a scheduled CDR scan may run with limited options. If a manual scan includes credentials or deeper scanning rules, the system treats the manual scan as more valuable. The scheduled scan ends, the manual scan takes over, and the original queue is lost.
There is no prioritization system today. New scans always join the back of the queue unless they require greater depth, in which case they replace an active action. Because progress is stored only in memory, the Sensor cannot resume interrupted scans.
How can I fix this
You cannot prevent this behavior entirely, but you can control when it occurs and reduce the risk of losing progress. Use the following best practices:
-
Keep large Discovery Actions under practical limits.
The UI displays warnings when an action contains around twelve thousand targets. Treat this as a soft boundary. Smaller actions complete faster and reduce the chance of overlap between scheduled and manual scans.
-
Avoid launching manual scans during heavy or long-running scheduled scans.
If a scheduled scan is crawling through thousands of devices, wait before running manual actions. This keeps the queue stable and prevents accidental cancellation. If a manual scan must run, create a smaller target list so it completes quickly.
-
Split very large ranges into smaller discovery jobs.
Instead of scanning a broad subnet in one sweep, divide it into manageable blocks. Smaller scans reduce memory usage in the Sensor, shorten scan time, and avoid losing hours of progress if the queue resets.
-
Ensure the Sensor service remains stable.
Because scan progress lives only in memory, any service restart clears the queue. Avoid restarting the Sensor during active scans. If maintenance is required, pause scheduled scans first.
-
Review scan configurations for overlapping depth.
If your scheduled scan uses minimal options and your manual scan uses full options, the system will replace the scheduled one. Align scan depth settings to reduce unintended overrides.
-
Educate users on queue behavior.
Teams often launch manual scans without realizing they can interrupt other jobs. Communicating this behavior prevents accidental disruptions.
This behavior is expected today, but it can be managed. By tuning your Discovery Actions and planning manual scans carefully, you can keep your scanning workflow steady and predictable.
