When the Raynet One agent is installed and enrolled on a device, inventory jobs targeting that device use the agent to perform the scan instead of agentless methods. The agent performs the inventory locally on the device and uploads the results to the platform through the runner.
How agent-based inventory works
When the server creates an inventory job for a device on which the agent is installed, the job instruction is queued by the server and picked up by the agent at the next sync (by default every 10 minutes). The agent then performs the local scan and uploads the results to the platform through the configured serving endpoint. This model relies solely on outbound connections from the agent, so no inbound network connectivity to the managed device is required.
Inventory jobs initiated by the server and those triggered manually using the inventory CLI command follow the same execution path on the agent. The difference is in how the job is initiated: server-side jobs are queued on the server and picked up by the agent at the next sync (default: every 10 minutes), while CLI-triggered jobs run immediately on the device without waiting for a sync cycle.
Duplicate inventory jobs
If an inventory job is already running or a completed job for the same device is pending upload, the agent skips a newly received job that would result in a duplicate run. This prevents unnecessary scans when multiple jobs for the same device arrive in the same sync cycle or when a manual inventory command is issued while a server-initiated job is in progress.
When a job is skipped as a duplicate, the outcome is visible in the job status in Raynet One, in the agent log, and in the failure report if available. A skipped duplicate job does not result in missing inventory data — the device's inventory is covered by the job that is already running or has already completed.
|
Note: |
