Sidekick volumes-from vs. convoy volumes

I try to build a service with two primary containers which use LBs and a shared persistent volume, but that isn’t supported by sidekicks:

So it looks like I have to use convoy instead of sidekicks / volumes-from?
Is there a way to share a volume between two primary services by sidekicks / volumes-from inside a stack?

I testing some other and simpler applications / stacks with just one primary and one sidekick (data only container). That should work fine.

But what should I prefer? mixed usage? Sidekick if possible? Is there a benefit to use sidekick volumes-from or convoy volumes? Performance? Backup / restore / move stacks?

When I need to use the same datacontainer for 2 containers, I ended up “duplicating” the datacontainer with a different name such as “html1” & “html2” both pointing to the same image… Its not the “cleanest” of setups but it gets the job done… In this case, since its “application files” which I dont want to get mixed up with the “server” containers… I dont think that a shared volume would be ideal in this scenario - at least I believe that its overkill?

Another question however, is if we can do “sidekick upgrades” in lieu of an entire service upgrade? or change the sidekick image without changing the service image and have that be more efficient? Using local docker-compose files with the amazon key, a upgrade of the service isnt upgrading the files/sidekick container… (I didnt try upping the datacontainer itself yet, but it could be I’m missing that…I’ll try to do that and post here… I guess I’ll have to update both sidekicks independently as well…)