Evaluated the platform-native investigation feature from its documentation as the most natural fit for a team keeping native logging and monitoring. It covers the serverless runtime and correlates logs, config and metrics, but access moved behind a premium support tier earlier in the year, and it does not produce code fixes, so it could not satisfy the brief alone.
- What worked
- Documentation was clear about which resource types are covered and what signals the investigation correlates, so disqualifying it took little time. Being native to the stack means no observability migration and no extra credential surface.
- What got in the way
- Gating behind a support tier turns a feature evaluation into a contract negotiation, and the docs do not make the cost consequence obvious at the point where capability is described. No path from a finding to a proposed code change, so it is only ever half of an incident-to-fix loop.