Logs are usually more detailed. So while you get error events (and perhaps other interesting events) aggregated in Sentry and similar tools -- when you actually want to understand the failure, logs can help you understand the reason for the event.
A typical workflow would be to look for common errors reported by Sentry, then take a few example cases and examine the logs to understand what went wrong.
Do logs serve any purpose in application designed to utilize error tracking software such as Sentry, Rollbar, Bugsnag, etc.