beacon_skill

Overview

This skill is an observability and reliability engineering specialist that designs SLOs, alerting strategies, distributed tracing, dashboards, capacity plans, and toil-reduction roadmaps. It focuses on strategy and specification; implementation is handed off to engineering agents. The goal is to make services measurable, actionable, and aligned with business outcomes.

How this skill works

Beacon inspects service behavior, business metrics, incident postmortems, and infrastructure context to recommend SLO/SLI definitions, error budgets, burn-rate alerts, tracing scopes, and dashboard views. It produces clear implementation specs, runbooks, escalation policies, and capacity models for downstream teams. It triggers stakeholder questions when SLO targets, sampling, or escalation policy choices require human input.

When to use it

• Defining SLOs and SLIs for user-facing APIs or services
• Designing an alerting strategy to reduce noise and focus on symptoms
• Specifying distributed tracing and OpenTelemetry instrumentation
• Creating audience-specific dashboards (dev, ops, biz)
• Planning capacity, autoscaling, and growth forecasting

Best practices

• Start with SLOs before adding new alerts or dashboards
• Define error budgets and burn-rate alerts to guide response
• Alert on symptoms not root causes and attach runbooks to every alert
• Use RED for services and USE for infrastructure metrics
• Design tracing sampling to balance observability and cost
• Prioritize toil by automation score and aim for self-healing where safe

Example use cases

• Design SLIs for an API to measure availability and latency and calculate a one-week error budget
• Create an alert hierarchy and runbooks to reduce pager noise during deployment windows
• Specify OpenTelemetry span naming and sampling for a microservice mesh
• Produce Grafana dashboard-as-code templates for SRE, dev, and product audiences
• Model resource needs and autoscaling policies for anticipated traffic growth