Octopus 2023.4 introduces the Resilient Tentacle Communications feature, designed to enhance communication reliability over unstable networks by automatically recovering from transient errors, ensuring successful deployments and runbook runs that might otherwise fail. This feature addresses issues faced in environments where Tentacle operates over complex or unreliable networks, such as corporate firewalls, satellite networks, or intermittent cellular connections, which can account for a notable percentage of failed deployments. The solution involves retrying remote procedure calls (RPCs) in an idempotent manner, ensuring that scripts aren't executed multiple times, with significant changes made to both Octopus Server and Tentacle to maintain backward compatibility. The implementation includes a new ScriptService and CapabilitiesService to determine Tentacle's supported features, allowing Octopus Server to handle communication more effectively. While the retries focus on communication errors, they might increase deployment duration if execution issues arise, but the overall success rate improves significantly. The feature is available by default on Octopus Cloud and can be enabled by updating both Octopus Server and Tentacle, with configuration options available through Machine Policies.