Typical starting point
Incidents recur; alerts are noisy; shared services lack clear owners; teams rebuild the same foundations; investigation depends on scattered information.
Focus area · Platforms & reliability
Run shared platforms and critical services with clear ownership.
Platforms and operational signals create value only when someone owns them and people can act on them. The work connects platform strategy, service ownership, telemetry, and incident learning so that leaders see how critical services behave and teams can act with confidence.
What the work covers
Service objectives, telemetry, incident response, and learning from failures, together with the platform service model that keeps shared capabilities adopted and maintained. Automation and AI support investigation or response where their actions can be checked.
How progress is judged
Service reliability, recovery time, recurring incidents, alert usefulness, platform adoption, and the effort required to operate the platform.
Platform engineering
Shared platforms should make it easier for teams to build, deliver, and operate software. The work spans platform architecture, self-service capabilities, delivery pipelines, standards, and the service model that keeps those capabilities useful.
It starts with the teams using the platform: where they wait, what they repeat, and what they need to own. Adoption, provisioning time, operational reliability, and maintenance effort show whether the platform is paying for itself.
Developer experience
A platform succeeds when developers choose to use it. That depends on the everyday conditions of development: getting started, understanding the codebase, running and testing changes, getting feedback, and releasing with confidence.
Related thinking
Signals, telemetry & bounded AI.
Perspectives on turning telemetry from passive dashboards into an active control plane for investigation, decisions, and operational safety.
Observability After Dashboards
Observability is becoming the control plane between software intent and production action—built on open signals, governed data, and bounded AI.
Follow the money - the future of observability
Agentic systems do more than generate telemetry. They reason across context and act in operations—changing what the reliability architecture must provide.
Built in public agent-otel-bridge (`v0.6.0`) streams sub-millisecond OpenTelemetry spans from coding agents in flight, and Dendro (`v0.7.0`) turns local JSONL transcripts into DuckDB fleet intelligence and a 60fps Ratatui cockpit at rest. Explore agent-otel-bridge Explore Dendro
Start a conversation
Describe the platform or reliability decision in front of you, what has been tried, and what is at stake.