Jaeger Mastery (Senior → Principal)

Operate

• Start from the incident and debugging questions traces must answer.
• Treat collectors, agents, query path, storage backend, and sampling as a unified system.
• Prefer operational clarity over trace feature sprawl.
• Design for trustworthy traces during outages, not only normal days.

Default Standards

• Sampling must reflect business-critical workflows.
• Storage and retention must match actual debugging value.
• Multi-tenant and access boundaries should be explicit where shared.
• Query UX should align with real investigation paths.
• Missing or partial traces must be diagnosable.

References