Cron Triggers in Hasura is not working locally when I created it via metadata API.

### Version Information

Server Version: 2.16.1
CLI Version (for CLI related issue): 2.16.1

### Environment

OSS / Cloud

### What is the current behavior?

- Cron Triggers in Hasura is not working locally when I created it via metadata API.
- On the Cloud version it's working correctly though.
- The event is getting created successfully but the pending events list is empty.
- When I try to delete the event it throws an error like it does not exist.
- If we create the event via the console it works. I actually need to use it via API only.
- The first time I create it via API and deleted it via API things are fine.
- In the same instance if I go a second time then pending events do not populate.

https://hasura.io/docs/latest/api-reference/metadata-api/scheduled-triggers/#metadata-create-cron-trigger

### What is the expected behavior?

- The pending events should get populated irrespective of how i create the event
- The cron events should fire as provided in the schedule

### How to reproduce the issue?

1. Spin up a Hasura instance locally using docker-compose
2. Call metadata API to create a cron trigger & using API it self delete it
3. Repeat 2 steps in the same instance. Now you can see that the pending events is not getting populated and the events are not getting fired

### Screenshots or Screencast


### Please provide any traces or logs that could help here.

### Any possible solutions/workarounds you're aware of?

- I use the console to create one more dummy trigger to make failed Tiggers load
- Use the cloud version but in that case lot of port forwarding needs to do

### Keywords

cron

Start by reproducing the issue with a local docker-compose Hasura instance through the metadata API's create and delete cron-trigger operations. Compare the repeated API-created trigger with a console-created trigger, then verify that pending events populate and cron events fire after the second create/delete cycle.

Written by the indexing model from the issue text.