USE CASE · KUBERNETES TROUBLESHOOTING
See workload, runtime, and change as one system.
Kubernetes failures cross cluster and app boundaries. Troubleshooting should not force you to rebuild that map by hand.
Kubernetes failures span workload, node, network, and deployment change—but tooling still forces engineers to hop between cluster and app views.
02 · AutoObserve
How AutoObserve addresses it
Topology and Investigation keep runtime entities, dependencies, and recent changes in one model tied to the active explanation.
CHECKOUT INCIDENT · SYSTEM LENS
Blast radius and deployment context stay attached to the explanation.
checkout-api v2.14.7 sits in a dependency path with payment and orders. Topology keeps that graph available while Investigation explains the regression—not as a separate cluster scavenger hunt.
03 · Resolution
What changes operationally
Troubleshooting answers 'what in the system changed?' without rebuilding the map by hand.