Paths by role

If you are not sure where to start, the maturity self-assessment places your organization on the four dimensions and suggests an action plan.

The site is open and can be read in any order. The paths below simply suggest a reading order for your role, across the four dimensions. They are not only for technical teams, since observability is worth little if business lines, finance and compliance do not know what to expect from it. Chapters not yet published are flagged in each overview.

New to observability, whatever your role? Start with observability in 10 minutes. Steps marked “coming soon” are not written yet: they will arrive as the site grows.

Engineer, SRE, developer

The aim is to instrument, collect and troubleshoot on your own.

1. Technical: signal fundamentals, then OpenTelemetry end to end.
2. Technical: telemetry pipelines, with the animated Collector pipeline until the Collector lab is out (coming soon).
3. Technical: SLOs and alerting.
4. People: troubleshooting under pressure (coming soon); meanwhile, the blameless post-mortem.
5. Technical, going further: observing an LLM system.

Matching courses, for generative AI, in order: AI in cross-section, understanding generative AI observability, the LLM labs, the deployment playbooks and VictoriaMetrics for LLMs, then the GenAI method and beyond the LLM.

Architect

This path is about how to choose a stack, size it and make it scale.

1. Technical: telemetry pipelines (the animated Collector pipeline), then backends and storage (coming soon).
2. Business: cost is decided at the Collector (Collector cost simulator), then pricing models (coming soon).
3. Business: sovereignty and architecture choices (coming soon); meanwhile, the solution scorecard.
4. Organization: telemetry governance and data contracts.
5. Technical: log integrity if you operate in a regulated environment.

Matching courses, for generative AI: start with understanding generative AI observability, then the deployment playbooks and VictoriaMetrics for LLMs; the GenAI method comes next, once the guide is under your belt. HPC AI reference architecture if you run GPUs.

Team manager

This path covers two concerns: making on-call sustainable and driving adoption.

1. People: alert fatigue, then sustainable on-call; the leadership view is in the CIO path, lesson 5.
2. People: blameless post-mortems.
3. Organization: SLOs as a contract between teams, starting from SLOs and alerting.
4. Organization: team models; roles and skills to drive adoption (coming soon).

Matching tools: the RACI matrix and the alert fatigue simulator.

Business line or product owner

You want to know what you can ask of observability, and get it.

1. What the business can ask of observability: the questions to ask your tech team and a checklist to bring to meetings.
2. Bridges: the translation dictionary between technical signal and business indicator.
3. Organization: SLOs as a contract between teams, phrased in your customers’ words: SLOs and alerting, skipping the formulas.
4. People: blameless post-mortems, where you belong too.
5. For AI: AI in cross-section, then understanding generative AI observability, chapters 1 and 3.

Finance, risk, compliance

This path helps you track a unit cost and hold evidence that stands up.

1. Bridges, section “To support functions”.
2. Business: the business case, then observability FinOps.
3. Technical, without reading code: a signed log is not an auditable log, sections 1 to 3.
4. For AI: evidence matrix and exposure calculator.

Leadership, CIO, procurement

The aim is to make informed trade-offs on budget, risk and sovereignty.

1. The course Governing observability: the CIO path, written for this profile: business case, compliance, governance, FinOps, people, steering.
2. Organization: the maturity self-assessment, then team models.
3. To prepare an executive committee meeting: the toolkit and the executive dashboard.

To get your bearings without a technical background: understanding high performance computing. To put figures on it: the GenAI exposure calculator.

Revised on 4 October 2026: generative AI steps aligned with the progression (AI in cross-section, guide, labs, production, method, beyond the LLM); the architect goes through the guide before the method.