DevOpsInterviewPrep logo
← Observability, SLOs & Reliability / 39
hardNewDatadogMicrosoftUber

Design the OpenTelemetry collector topology for four hundred services. Where does sampling happen?

Agent per node, gateway per cluster, and one decision that shapes everything else: tail sampling needs every span of a trace in one place, which is why the topology has a tier that looks redundant until you need it.

Updated Sep 2026 · Grounded in researched DevOps, SRE and platform engineering interview loops, written to a senior-engineer editorial bar, and never padded to hit a word count.

Agent per node, gateway per cluster, and one decision that shapes everything else: tail sampling needs every span of a trace in one place, which is why the topology has a tier that looks redundant until you need it.

20 answers per topic instead of 10, and your progress kept · no cardor unlock all 390 remaining answers · ₹2,000 / $25
UP NEXT ON YOUR JOURNEY
Next in this trackLatency regressed 30 percent and traces show the time is inside one service. Now what?Next in this trackYour architecture is queues and consumers. The golden signals assume request and response. Now what?Next in this trackYour OTLP endpoint returns HTTP 200, but spans are missing. Should the Collector retry the batch?
DISCUSSION · 0

Nothing here yet. Say how you would answer it.