[ENG-6052] | Fixes to share for M chips discovered during local setup

[sh-andriy]

No description provided.

[sh-andriy]

Were encountering errors when running this command docker-compose up -d elastic8, such as network problems (share_network helped), still had other issues and after specifying platform it removed incompatible platforms warnings (arm with amd) and set elastic8 up without troubles.

When was testing the share project in osf (here http://localhost:5000/search) could not see any of my projects and encountered errors this fix helped me to solve them:
in docker-compose.yml:

in elastic8 service under environment category removed this line:
- cluster.initial_master_nodes=singlenode
add this line instead:
- discovery.type=single-node

[mfraezz]

Networking changes should be fine, and the change from cluster.initial_master_node to discovery.type makes sense -- if we're only running a one-node cluster, it should elect itself whenever it spins up.

I personally ran into an issue spinning up rabbitmq:management when it attempted to enable the stream_sac_coordinator_unblock_group feature flag, a recent addition in version 3.13. I found that specifying the 3.12-management-alpine tag instead allowed it to run. This may be a hardware issue, as I found reports of other M-series apple chips encountering the same or similar issues.

Overall, LGTM, but I'll leave it for @aaxelb to confirm and merge.

[aaxelb]

looks good, thanks! (and test failure unrelated)