Project Description
Our real-time data platform needs a rock-solid proxy layer that can move huge volumes of HTTP and SOCKS5 traffic with minimal latency. The immediate focus is designing and automating a high-concurrency architecture—from the Squid clusters themselves to the surrounding network and IP allocation logic—so that new nodes come online in seconds and stay performant under sustained load.
Key components of the engagement
• A web-based proxy panel that provisions, monitors, and rotates IPs automatically, exposing a clean API for our backend services.
• Fully scripted Squid deployments on Linux, container-ready (Docker) but not container-bound, with Redis or similar caching where it measurably improves throughput.
• Traffic distribution and fail-over policies tuned via Nginx, LVS, or your preferred load-balancing stack, continuously benchmarked for connection time and packet loss.
• Infrastructure-as-code (Ansible, Terraform, or equivalent) so the entire fleet can be recreated from scratch and scaled horizontally without manual intervention.
• Performance dashboards and alerting hooks that let us visualize connection counts, CPU, memory, and per-endpoint latency in real time.
Acceptance criteria
1. Spin-up script launches a new proxy node that registers itself in the panel, reachable within 5 minutes.
2. Sustains at least 50 000 concurrent connections with <100 ms added latency on a standard 1 Gbps link.
3. Automated health checks gracefully drain traffic and recycle failing nodes without dropping active sessions.
4. Clear, maintainable documentation covering architecture decisions, configs, and rollback procedures.
Work style
Remote, long-term collaboration with a flexible schedule. You’ll be interacting directly with senior engineers in a highly technical environment; expect code reviews, performance tests, and room for architectural input. If tuning kernels, squeezing extra RPS out of Squid, and scripting zero-downtime rollouts sound exciting, let’s get this running.