Closely related to the previous strategy, you may try optimistically to deploy your code to production but, for some time, only expose part of the traffic to the new code.

You can build a system where only a small percentage of the traffic will hit the new servers running the new code and compare the error rate and performance originating from each release for a short period. Usually, 10% of the traffic for 10 minutes is enough to collect enough signal about the quality of the new build. If after that time, everything looks good, move 100% of the traffic to the new version of the service. Bugs such as memory leaks are usually slower to manifest themselves; once the deployment is completed, continue closely monitoring your different system and key metrics to make sure everything is good: