About
I run production infrastructure for a living: servers, deployments, monitoring, on-call rotations. I also write software, mostly backend. This blog is where I write about both sides of that work — the two posts up now, on logging and message queues, are a fair sample of what I spend my days on.
Blog
Dynamic mapping is convenient right up until it isnt. Field explosion, ILM policies, shard sizing, and the index template that keeps an Elasticsearch cluster from mapping itself into a corner.
ELK is four separate tools that happen to ship together. What Elasticsearch, Logstash, Beats, and Kibana each actually do, how they fit into a pipeline, and when the stack is worth the operational weight.
Kafka and RabbitMQ solve different problems: replayable event streams vs routed task queues. A breakdown of both mental models, when to use each, and why some architectures need both.
Monitoring tells you WHEN something broke. Observability tells you WHY. A practical guide to structured logging, correlation IDs, intentional log levels, and connecting logs to traces.