Metrics collected properly, analysed properly, explained honestly.
I build analytics and observability systems for organisations that need more than a dashboard. The work covers collection, federation, consolidation, storage, visualisation, deep analysis, and reporting — with a focus on technical accuracy rather than narrative spin.
I help teams find hidden value in data that appears mundane, and shape collection and storage so that analysis stays defensible when it matters.
Most organisations have scattered metrics — some in Prometheus, some in application logs, some in database tables, some nowhere at all. Before a dashboard is worth building, the collection layer needs to be honest about what it is actually measuring.
I design and build the full pipeline: collection and federation, storage in VictoriaMetrics or Prometheus-compatible systems, Grafana visualisation built for the people who will read it, and deeper analysis beyond what any dashboard panel can show.
Fast, efficient time-series storage. Strong ingestion performance, efficient on-disk storage, and practical querying at scale.
Dashboards, exploration, and custom visualisations. Built for engineers, not sales decks.
Custom collector for MySQL, PostgreSQL, MSSQL, MongoDB, Elasticsearch, Kafka, CSV, AWS Athena and more. Hot-reload config, concurrent workers, retry logic.
TUI for real-time worker and pool monitoring + remote TUI access.
Local (Ollama, llama.cpp) or hosted models that turn raw time-series data into plain-language explanations, anomaly detection, and actionable reports.
Home monitoring covering power, temperature, humidity, chicken coop, UPS, and Starlink in a rural location with unpredictable mains voltage. Real instability generating real data worth analysing.
Correlating voltage drops • Brownout risk analysis • DeepThought explanations
Available for remote and async consultancy across collection, analysis, and observability engineering.