circuit breaker pattern, overkill or essential

lately, i've been thinking about the circuit breaker pattern for handling failures in microservices. we're seeing more inter-service calls and the potential for cascading failures. is implementing a circuit breaker for every remote call just adding unnecessary complexity and overhead, or is it a crucial safety net we should be employing more broadly? what are your experiences with it? do you find it truly helps prevent outages or just adds another moving part to manage