Flyte enables production-grade orchestration for machine learning workflows and data processing created to accelerate local workflows to production.

Flyte

Hello!
I've been setting up Flyte via the <https://docs.flyte.org/en/latest/deployment/deployment/cloud_simple.html|single cluster deployment docs> and the flyte-binary service fails to start with the following error:
```2023/03/08 22:40:38 /go/pkg/mod/gorm.io/driver/postgres@v1.2.3/migrator.go:74
[0.326ms] [rows:1] SELECT CURRENT_DATABASE()
{"json":{"src":"start.go:158"},"level":"panic","msg":"Failed to start Admin, err: database migration failed: ERROR: relation \"description_entities\" does not exist (SQLSTATE 42P01)","ts":"2023-03-08T22:40:38Z"}
panic: (*logrus.Entry) 0xc000aa8460```
This is into a fresh RDS database.  The migration runs up through `2021-08-05-schedulable_entities_snapshot` verified via `select * from migrations order by id desc limit 1;`.  Any help is appreciated!

I wonder if this is a PG 15.2 issue
postgres logs:
```2023-03-08 22:51:26 UTC:10.3.51.212(37283):user_flyte@flyte:[2988]:ERROR: relation "description_entities" does not exist
2023-03-08 22:51:26 UTC:10.3.51.212(37283):user_flyte@flyte:[2988]:STATEMENT: ALTER TABLE IF EXISTS description_entities ALTER COLUMN "id" TYPE bigint```
attempting `ALTER TABLE IF EXISTS description_entities ALTER COLUMN "id" TYPE bigint` locally results in an error which shouldnt be the case with the `IF EXISTS` condition

attempting that command in a different PG 14 db results in no error, will downgrade PG and reattempt

good catch, we should update the command to support PG 15

downgrading to PG 14 did resolve the issue.  I don't see any explicit mention in PG 15 release notes over changes to `IF EXISTS`, but there are changes to default permissions which I believe could effect its behavior