Automation
Prompt engineering
Context
Customer wanted to gain better visibility into their information system. Until then, no centralized monitoring solution was in place, which complicated incident analysis and slowed down decision-making.
Problem
The IT system suffered from sporadic latency, slowdowns, and opaque overall activity. It was impossible to quickly identify bottlenecks or the most resource-intensive applications.
Solution
I designed and deployed a complete observability stack:
- Umami to collect analytics and track user behavior in real time.
- Prometheus to aggregate system and application metrics.
- Loki + Alloy to centralize and index logs from all services.
Impact
- Detection and removal of unnecessary processes, freeing up 15% of RAM.