Overview 💥
The Liongard Windows Agent is designed to operate quietly in the background with minimal system impact. Under normal conditions, the agent performs lightweight tasks such as:
Retrieving current inspector configuration
Triggering inspections according to schedule
Sending collected data back to the Liongard platform
Most of the time, the agent remains idle and consumes negligible CPU. However, temporary CPU spikes may occur when several inspectors execute at the same time—especially when they share the same agent.
Liongard includes scheduling tool—such as Distributed Start Times and the Spacing feature—to help prevent these overlaps and ensure smooth, predictable performance.
Why You May Notice Higher CPU During Inspections 🤔
This may happen when :
1️⃣ Multiple Inspectors Linked to the Same Agent
When several inspectors—Windows, Active Directory, Hyper-V, SQL, VMware, etc.—use the same agent, they all rely on that same server to execute inspection tasks.
If they all run at once, you get a stacked workload.
[00:00] ─────► Windows Server Inspector Run
[00:00] ─────► AD Inspector Run
[00:00] ─────► Hyper-V Inspector Run
[00:00] ─────► SQL Inspector Run
│
▼
┌──────────────────────┐
│ SAME AGENT BUSY │
│ Tasks running in │
│ parallel = load spike│
└──────────────────────┘
2️⃣ All Inspectors Scheduled at the Same Time
When everything is set to 00:00 or Every hour at :00, the agent receives multiple jobs simultaneously.
3️⃣ Distributed Start Times Disabled
If staggered scheduling is not enabled, all linked inspectors cluster together by default, increasing the chance of simultaneous workload bursts.
How to Reduce CPU Spikes During Inspections 🧑🏫
The most effective and supported method to prevent CPU spikes is to space out inspector launchpoints so they do not run at the same moment. When many inspectors are linked to the same agent and their inspections overlap, the agent processes multiple workloads in parallel — resulting in temporary CPU spikes.
Spacing solves this by distributing launches over a defined interval. Example Scenario:
You have 6 inspectors connected to one agent
Spacing = 10 minutes
Liongard will distribute the runs like this:
Inspector | Scheduled Start | Spacing Applied | Final Run Time |
Windows Agent | 00:00 | — | 00:00 |
AD | 00:00 | +10 minutes | 00:10 |
Hyper-V | 00:00 | +20 minutes | 00:20 |
SQL | 00:00 | +30 minutes | 00:30 |
Exchange | 00:00 | +40 minutes | 00:40 |
VMware | 00:00 | +50 minutes | 00:50 |
How to Apply Spacing in Liongard :
Follow these steps to stagger your inspector launchpoints:
1. Login to your Liongard platform and navigate to: Admin → Inspectors
2. Select the Inspectors / Launchpoints
Choose the Inspector type (e.g., Microsoft 365, Active Directory, etc.)
Check the boxes next to all the launchpoints you want to reschedule together
3. Open the Scheduling Menu
Click the Actions tab
Select Set Inspector Schedule
4. Configure Spacing
In the scheduling panel, choose: Space These Launchpoints Out By…
5. Enter Your Interval and Save
If spacing launchpoints: choose how far apart you want them scheduled
Example: 5 minutes, 30 seconds, 1 hour, etc.
Click Save to apply the configuration
Liongard will reorder the inspectors according to the spacing you set, preventing them from firing simultaneously.
Scheduling Logic 🌟
┌────────────────────────┐
│ Are multiple inspectors│
│ using the same agent? │
└───────────┬────────────┘
│ Yes
▼
┌────────────────────────┐
│ Are their start times │
│ overlapping? │
└───────────┬────────────┘
│ Yes
▼
┌───────────────────────────┐
│ Enable Distributed Start │
│ Time + Apply Spacing │
└───────────────────────────┘
When to Contact Liongard Support 🦁
Please reach out to Liongard Support if:
You consistently see very high CPU usage (e.g., > 50–70%) during inspection windows.
You suspect that inspections or directories are misconfigured (leading to excessive work).
Newer Agent versions don’t improve performance after upgrade.
You want help tuning inspector frequency, scheduling, or workload.
Helpful details when submitting a ticket:
Agent version (e.g., 5.x).
System specs: CPU, memory, OS version.
Your inspector schedule (which inspectors run when).
Observed CPU usage (with screenshot or monitor data).
Log files or recent inspection run duration.