Overview
mxsec-platform is the MatrixPlus open-source Cloud Workload Protection Platform (CWPP), purpose-built for the cloud-native era. It provides a unified security plane across hosts, containers, Kubernetes, and serverless workloads, and consolidates runtime detection, vulnerability management, baseline compliance, container security, and antivirus into a single control plane — so DevSecOps teams can run real defense-in-depth, not just dashboards.
mxsec-platform is currently deployed on 3000+ production hosts, spanning overseas subsidiaries of multinational groups, government and enterprise customers, and internet-business workloads.
Key metrics
- 3000+ production hosts
- AGPL-3.0 open-source license (Community Edition)
- Two editions Community + Enterprise
- 8 Kafka topics + DLQ for business-domain data-plane partitioning
- 6 schedulers canary rollout / rule sync / alert governance / IOC sync / plugin distribution / restart orchestration
- Low overhead eBPF-based runtime detection, typically under 2% CPU and 80 MiB RAM
- mTLS everywhere end-to-end Agent ↔ Server encryption, with first-touch enrollment
Architecture overview
mxsec-platform separates control plane / data plane / detection plane into three independently scalable tiers — no single bottleneck.
graph LR
subgraph WL[Workloads]
H[Host]
C[Container]
K[Kubernetes]
S[Serverless]
end
AG[mxsec-agent
eBPF · Plugins · Local detection]
H --> AG
C --> AG
K --> AG
S --> AG
AG -->|gRPC BiDi · mTLS| AC[AgentCenter ×N
Ingestion / command dispatch]
subgraph CP[Control plane]
MGR[Manager ×N
REST API · JWT · RBAC]
REDIS[(Redis
Service discovery · cache)]
MYSQL[(MySQL
Rules · tasks · assets)]
end
AC <-->|HTTP control| MGR
MGR <--> REDIS
MGR <--> MYSQL
AC -->|8 topics + DLQ| KAFKA[(Kafka
Business-domain data plane)]
KAFKA --> CONS[Consumer ×N
CEL · sequence detection · enrichment]
subgraph DP[Detection & storage]
ENG[Engine
Alert governance / correlation]
CH[(ClickHouse
Events · metrics)]
end
CONS --> ENG
CONS --> CH
ENG -.->|Alerts| MGR
MGR -->|HTTPS| WEB[Web console
SPA + REST API]
ENG -.->|Forward| SIEM[Syslog / Kafka
SIEM / ticketing]
Core capabilities (seven modules)
1. Server architecture
- Three independently scalable tiers:
manager/agentcenter/consumer - Embedded service discovery: Redis HSet + Pub/Sub + 30s reconciler
- Active health probing (10s probe + 60s heartbeat TTL) plus failure counter
- Power-of-Two-Choices load balancing — no hot-AC tail latency
- Reliable command delivery: MySQL persist → Redis routing → AC HTTP → gRPC → Agent ACK
2. Agent
- Single-process Go agent + plugin lifecycle + watchdog
- Backoff restart and sleep/wake state machine
- WAL + in-memory cache for 7-day offline buffering
- Resource ceiling: under 2% CPU and 80 MiB RAM
- mTLS everywhere, with first-touch enrollment (Server issues client cert)
- Canary upgrade: 5% → 25% → 100% with failure pause and auto-rollback
3. Baseline compliance
- Linux + Kubernetes dual rule model
- CIS / NIST / China MLPS 2.0 baselines out of the box
- Canary rule rollout with hot reload — no Agent restart
- Remediation loop: scan → recommend → canary remediation → verify
4. Runtime / EDR
- eBPF-based real-time behavior collection and anomaly detection
- Attack-chain (Storyline) visualization
- Port-scan / web-shell / reverse-shell detection out of the box
- CEL engine + sequence detection + correlated alert governance
- One-click block / isolate / file quarantine response
5. Container / Kubernetes
- Three-tier view: node / pod / cluster
- Image vulnerability + secret + Dockerfile scanning
- K8s audit log ingestion (including GCP Pub/Sub)
- CIS Kubernetes benchmark for cluster baselines
6. Vulnerability management
- Most comprehensive multi-source advisory ingestion in the Chinese open-source ecosystem (GitHub / NVD / domestic vendors)
- Unified view across host / image / source code
- CVE priority with EPSS feeds
- One-click ticket handoff and remediation guidance
7. Antivirus / IOC
- YARA-X rule engine + IOC reputation feeds
- File, process, and domain reputation
- Auto-updated virus database with hot reload
Integration with the MatrixPlus stack
mxsec-platform is not an island — it integrates natively with the other three MatrixPlus open-source products:
| Integration | Value |
|---|---|
| mxcmdb · asset graph | Alerts and vulnerabilities link directly to assets; blast radius is visible at a glance |
| mxid · unified identity | Console SSO, operation audit, API authentication |
| ticketdesk · ticketing | One-click alert-to-ticket, full remediation loop |
Deployment options
| Form | Best for |
|---|---|
| Binary | Traditional hosts, single- or multi-node direct install |
| Docker / Compose | Single-node, container-first, out-of-the-box |
| Kubernetes (Helm) | Multi-cluster, multi-AZ, highly available |
| Air-gapped | Government / regulated environments, fully offline |
Where it fits
- Large-scale Kubernetes security governance
- Unified security across multi-cloud and hybrid-cloud
- Transition-phase protection while migrating to cloud-native
- China MLPS 2.0 compliance and regulatory audit
- Remote security operations for overseas subsidiaries
- Red-team / blue-team exercise support
Live deployments
- Japan subsidiary of a UK-listed group — mxsec-platform Enterprise + 2 on-site engineers, one year, zero breaches
- Government and enterprise customers — security operations support during national-level red-team exercises
- MatrixPlus self-operated business — full-stack protection for our in-house commerce system
Get started
→ Repository: github.com/matrixplusio → License: AGPL-3.0 → Documentation and deployment guides ship with the repo
For private deployment, custom integration, Enterprise Edition licensing, or long-term operational support, contact us.