add back publish under original cluster name for symlink cluster

[bohhyang]

In a previous PR, we removed publishing to the original cluster name (e.g: FooCluster-prod-ltx1) for a symlink cluster ($FooCluster). But there could be edge cases that the original cluster was subscribed by direct curli calls (like calling the colo-suffixed service: d2://FooService-prod-ltx1) or hard-coded custom code.

This change add this publish back.

Test Done

Unit tests.
Tested with QEI d2-proxy,
call d2 service "accountBalances" (which is under cluster "$OmsServiceMaster"):
curli -v "d2://accountBalances" --d2-proxy-url "http://localhost:21360/d2/" -X OPTIONS --force-insecure-d2
HTTP/1.1 200 OK
< Content-Type: application/json
< x-linkedin-processing-colo: ei-ltx1
< x-linkedin-processing-machine: ltx1-app6705.stg.linkedin.com
(was able to pick the host ltx1-app6705.stg.linkedin.com).

, and call d2 service "accountBalances-ei-ltx1" (which is under cluster "OmsService-ei-ltx1"):
curli -v "d2://accountBalances-ei-ltx1" --d2-proxy-url "http://localhost:21360/d2/" -X OPTIONS --force-insecure-d2
HTTP/1.1 200 OK
< Content-Type: application/json
< x-linkedin-processing-colo: ei-ltx1
< x-linkedin-processing-machine: ltx1-app7220.stg.linkedin.com
(was able to pick the host ltx1-app7220.stg.linkedin.com).

[PapaCharlie]

Thank you for fixing this! What a weird and frustrating bug...