It looks like we have some news regarding this issue.
We are fairly confident that the issue relates to the way in which Octopus server is establishing connections to Tentacles associated with the deployment. Typically, these connections should happen in parallel, however it looks like they are currently happening sequentially. This can cause some serious delays if there are machines which Octopus cannot establish a connection with during the deployment.
The following GitHub Issue has a more detailed explanation of this issue and will have updates on the release for the fix.
Currently this is affecting Octopus version 2019.5.2 - 2019.5.10 (Inclusive). We have a fix planned for the 2019.5.11 release which we will hopefully release today or tomorrow.
The current workaround outside of upgrading to 2019.5.11 when it is released is to downgrade your Octopus server version to an unaffected version.
Alternatively, if you take action to ensure that Octopus can quickly connect to all machines, this issue should not be encountered. This could mean deleting any inaccessible machines.
The best option here would be to upgrade to our latest version of Octopus Server 2019.5.11 when it is released.
If after the upgrade to 2019.5.11 and are still encountering performance issues, please let us know.
Let me know if you have any questions or thoughts here.