Week 3’s instrumentation clinic caught that our “user happiness” gauge was really a sampling artifact. Painful, then clarifying.
Yuna · platform engineer
Flagship course
Program Metrics Lab
A six-week cohort where application analytics finally serve an observability program — with owners, cadence, and narratives that survive finance review.
Learning outcomes
- Define program metrics that roll up service and product signals without hiding trade-offs.
- Audit application analytics for decision value, cardinality cost, and ownership gaps.
- Facilitate a critique that improves definitions instead of defending dashboards.
- Deliver a fifteen-minute narrative executives can fund without requiring a live demo of every panel.
Modules
- Week 1 · Map the fog Inventory existing telemetry and the decisions that stall without evidence.
- Week 2 · Definition cards Write program metric contracts: owner, refresh, blind spots, related SLOs.
- Week 3 · Instrumentation intent Align application analytics to those contracts; retire low-value counters.
- Week 4 · Scorecard assembly Build a review-ready program view with narrative anchors.
- Week 5 · Stakeholder rehearsal Present under timeboxes; receive structured peer critique.
- Week 6 · Handoff ritual Codify cadence, escalation paths, and a 90-day improvement backlog.
Instructor
Sora Han
Sora spent a decade bridging SRE and product analytics teams across Korea and remote hubs. She facilitates the Lab with a soft voice and firm insistence that every metric name a decision. Outside cohort weeks she hosts metric tea sessions from the Daejeon studio.
Informational pricing
Individual Lab seats are typically purchased as Metric Atelier (see pricing). Team cohorts use Observatory Circle sponsorship. No checkout runs on this website.
Learner notes
★★★★☆
Rehearsals were excellent. I wanted one more async office hour for SQL-heavy folks; the glossary add-on helped mid-cohort.
Anonymous client in payments
FAQ
Do I need an existing observability stack?
Yes — at least one production system with some telemetry. The Lab is not an introduction to installing agents from scratch.
Is the Lab tool-agnostic?
Mostly. We reference common patterns across popular platforms, but exercises focus on definitions and narrative, not vendor certification.
What is a real limitation of this course?
We do not provide hands-on production access to your systems. Learners must bring exports, screenshots, or sandboxed examples. If your security policy blocks sharing even redacted panels, progress will be slower — and we cannot waive that constraint.
How does this relate to application analytics?
We treat product and application analytics as inputs to observability program metrics, not a separate silo — so activation, reliability, and cost signals can share a review language.