I will build enterprise system and network monitoring solutions
Senior DevOps and Network Security Engineer and Linux Expert
About this Gig
Eliminate monitoring blind spots. I am a Senior DevOps & System Engineer specializing in Zabbix, Grafana, and advanced Docker networking. Whether you need a fresh production deployment or immediate troubleshooting for broken stacks, I will deliver a rock-solid observability pipeline.
What I Deliver:
- Native & Containerized Stacks: Deploying Zabbix Server, Agent 2, and Grafana using isolated networks or host mode.
- Advanced Troubleshooting: Fixing critical failures like Docker ContainerConfig errors, Agent crash loops, and "Connection refused/reset" faults.
- Routing & Isolation Fixes: Overcoming Docker bridge constraints, host network limits, and Calico/K8s NAT source IP masking issues.
- Custom Dashboards & Alerts: Creating high-density Grafana panels via Zabbix API and integrating real-time notifications to Slack/Telegram.
- Why Me? Tailored solutions for specific cloud topologies (Hetzner, AWS) and firewalls (UFW/iptables). Precise runtime patches for long-term container stability.
Tools:
Kubernetes
•
Docker
•
Amazon EKS
•
Google Kubernetes Engine
Frameworks:
Npm
•
Terraform
•
Ansible
Programming language:
Bash
•
Java
•
JavaScript
•
Python
•
Ruby
Expertise:
Installation
•
Debugging
•
Configuration
FAQ
How do you capture host-level metrics inside a container?
I deploy Zabbix Agent using host networking (--network host), privileged flags, and system volume mounts. This bypasses container abstraction, allowing accurate reading of host hardware, disk I/O, and network throughput.
Can you fix "Connection reset by peer" or access permission errors?
Yes. This happens when the agent drops connections due to ACL mismatches caused by Docker bridge routing or Calico/K8s NAT masking the server's real IP. I re-architect the environment access controls to fix the handshake.
Why does my Zabbix Agent container loop-crash with "failed to validate ServerActive"?
he official image mirrors ZBX_SERVER_HOST into ServerActive. Active servers reject CIDR blocks (like /12). I break this loop by overriding the entrypoint configuration or enforcing explicit environment variables.
Will this setup interfere with my production applications?
No. All monitoring components are isolated using independent container networks or specific host port allocations, ensuring zero cross-contamination with your existing stacks or databases.

