The shape of a system, which services exist, how they connect, where they run, the substrate every other observability technique builds on.
Topology is the shape of a distributed system, which services exist, which call which, which run in which region, which are public-facing, which are internal. Modern observability tools auto-discover topology from telemetry, traces show service-to-service edges, mesh sidecars enumerate connections, k8s API surfaces the workload graph. Knowing the topology is the prerequisite for everything from blast-radius calculation to dependency-aware alert correlation.
Most architecture diagrams in Confluence are wrong, drawn at launch, never updated. Live, auto-discovered topology is the only kind worth trusting during an incident, because the team's assumption of how the system is shaped is often weeks behind reality. Topology freshness, how recently the auto-discovery ran, is itself a metric worth tracking.
See the part of the platform that handles topology in production.