Build monitoring strategies

Use this guide to design or refine monitoring coverage with the Assistant’s help. It is ideal when you launch a new service, refactor architecture, or align with compliance or SLO goals.

What you’ll achieve

Describe a service, its dependencies, and business goals so the Assistant can propose the right signals.
Receive recommended dashboards, alerts, and runbooks tailored to your stack.
Produce a summary you can present to stakeholders or use as a planning checklist.

Before you begin

Service description: Summarize the architecture, dependencies, and critical paths.
Business objectives: Define the SLOs or KPIs you want to protect.

Describe the service and goals

Outline how the service works, what success looks like, and any constraints so the Assistant can target relevant observability signals.

We are launching a payments API. Suggest the metrics, logs, traces, and profiles we should collect.

Include deployment details (Kubernetes, serverless, multi-cloud) so the Assistant tailors its advice. Share examples of existing dashboards or alerts if you want the new plan to mirror them.

Request instrumentation and dashboard plans

Ask the Assistant to map metrics, dashboards, and alerts to your requirements, adjusting the prompts until the plan fits your stack.

Ask for recommended metrics and labels, along with sample queries.
Break the request into separate prompts (metrics, dashboards, alerts) if the response starts to drift; polite corrections keep the Assistant focused.
Request dashboard outlines or alerting strategies.

Recommend a dashboard layout for the payments API with latency, errors, and business KPIs.

Plan operational workflows

Translate the monitoring plan into day-to-day practices by exploring how on-call teams and collaborators will use the new coverage.

Ask how to integrate the strategy with incident response or runbooks.
Request collaboration tips, for example, How should the on-call team use infrastructure memory during incidents?.

Iterate and validate the plan

Pressure-test the recommendations by sharing limitations and asking for executive-ready summaries.

Share constraints (storage limits, noise thresholds) and ask for adjustments.
Request a concise summary you can take to project stakeholders.

You now have a monitoring strategy draft that covers signals, dashboards, alerts, and operational workflows. Document it with your architecture notes and iterate as implementation progresses.

Troubleshooting

Advice feels generic: add more detail about your stack or existing observability coverage.
Missing compliance considerations: mention regulatory requirements or data residency constraints.
Need implementation steps: ask for specific tasks, owners, and timelines.
Pivoting to a different service: open a new conversation so the Assistant discards the previous strategy context.