What is eBPF?
eBPF (extended Berkeley Packet Filter) is a revolutionary technology that allows programs to run safely inside the Linux kernel without changing kernel source code or loading kernel modules. AlertKick uses eBPF to provide deep, real-time security monitoring of your servers.
How eBPF Works
Section titled “How eBPF Works”Traditional security tools run in user space and can only see what the operating system chooses to expose. eBPF runs directly in the kernel, giving it complete visibility into system activity as it happens.
How eBPF monitors kernel events and streams them to AlertKick
The Key Components
Section titled “The Key Components”- eBPF Programs: Small, verified programs that attach to kernel events (called “probes”)
- Kernel Probes: Hooks at critical points like file access, process creation, network connections
- Event Streaming: Security events are captured and sent to the AlertKick agent in real-time
- AI Analysis: Events are analyzed, classified using MITRE ATT&CK, and alerts are triggered
Why eBPF for Security?
Section titled “Why eBPF for Security?”Complete Visibility
Section titled “Complete Visibility”eBPF sees everything that happens on your system:
- File operations: Every file opened, read, written, or deleted
- Process activity: Every command executed, including arguments
- Network connections: All outbound and inbound connections
- Privilege changes: sudo usage, capability changes, setuid calls
- Container escapes: Attempts to break out of container boundaries
Zero Blind Spots
Section titled “Zero Blind Spots”Traditional security tools have gaps:
| Traditional Tools | eBPF |
|---|---|
| Can be bypassed by rootkits | Runs in kernel, sees raw events |
| Miss fileless attacks | Sees all process execution |
| Log-based (can be deleted) | Real-time event capture |
| Agent overhead | Minimal CPU/memory impact |
Safe by Design
Section titled “Safe by Design”eBPF programs are:
- Verified before loading (can’t crash the kernel)
- Sandboxed (limited access to kernel functions)
- Read-only (observe only, no system modifications)
How AlertKick Uses eBPF
Section titled “How AlertKick Uses eBPF”Real-Time Threat Detection
Section titled “Real-Time Threat Detection”When suspicious activity occurs, AlertKick’s eBPF agent:
- Captures the event at the kernel level (no way to hide)
- Enriches with context (process tree, user, container)
- Streams to AlertKick cloud securely
- Classifies using MITRE ATT&CK framework
- Analyzes with AI to determine severity and context
- Alerts the right people through your escalation policies
Example: Detecting a Privilege Escalation
Section titled “Example: Detecting a Privilege Escalation”Event: setuid() calledProcess: /tmp/.hidden/exploitParent: /bin/bash (PID 1234)User: www-dataContainer: web-frontend-abc123
MITRE ATT&CK: T1548 - Abuse Elevation Control Mechanism
AI Analysis: Unusual privilege escalation attempt from webserver process. Binary location (/tmp) is suspicious.Recommend immediate investigation.
Severity: CRITICALWhat We Monitor
Section titled “What We Monitor”| Category | Events Monitored |
|---|---|
| File System | Access to sensitive files (/etc/passwd, SSH keys, configs) |
| Processes | Shell spawns, script execution, binary execution |
| Network | Outbound connections, DNS queries, port binding |
| Privileges | sudo, su, setuid, capability changes |
| Containers | Namespace changes, escape attempts, mount operations |
Getting Started
Section titled “Getting Started”Ready to deploy eBPF security monitoring?
- Install the agent: See eBPF Security Monitoring for setup instructions
- Create security profiles: Define what to monitor for your environment
- Configure alerts: Set up escalation policies for security events
Learn More About eBPF
Section titled “Learn More About eBPF”Want to dive deeper into eBPF technology?
Next Steps
Section titled “Next Steps”- eBPF Security Monitoring Setup - Install and configure eBPF agents
- MITRE ATT&CK Overview - Understand threat classification
- Escalation Policies - Configure alert routing