observability in microservices: metrics, logs, or traces?

hey everyone, im lookin to refine our observability strategy for our microservices architecture. we're currently collectin metrics, logs, and traces, but im wonderin what the communitys priorities are. when debuggin a complex issue across several services, which of these signals do you find most valuable, and why? are there specific tools or techniques you prefer for correlatin these different data points effectively