Wiring AI agents into GitHub, Jira, and alerting
An AI agent that can only chat is a demo. One that can read your repo, your tickets, and your alerts, and act on them, is a coworker. A practical look at the integration work that gets you there.
Analyzing Prometheus metrics with LLMs: the token problem
Pasting a /metrics scrape into a model sounds easy until you watch the token counter. The costs, failure modes, and shapes that actually work.
Why OTLP changes the observability stack
OpenTelemetry has been quietly reshaping observability for years. OTLP, its wire protocol, is the part that actually changes how engineering teams build telemetry pipelines.
LLM observability: an open-source toolkit
LLM features fail in different shapes than HTTP services. Here's a working open-source stack, OpenLLMetry, Langfuse, Phoenix, and how they fit together.
RAG that holds up: connecting org data to LLMs
Most production RAG systems get to demo, then quietly underperform. The model isn't the bottleneck, retrieval is. Here's what reliably moves the bar.