[BUG][SanityTest] EMR-S batch jobs automatically became success if no query submit for a while

[LantaoJin]

What is the bug?
The inner EMR-S batch job will completed as success when we don't trigger a Spark job for a while (seems it has a TTL in ~5 mins), then the next query will re-launch a EMR-S batch job which cause very slow.

How can one reproduce the bug?

1. Go to Workbench page. For example https://search-flint05-sanity-h5if7yelmxws5hc35cap2lauf4.us-east-1.es-integ.amazonaws.com/_dashboards/app/opensearch-query-workbench#/
2. Execute multiple same PPL queries work fine. For example, the following query after second run only takes seconds

source = myglue_test.default.nested | head 10

4. But when you hold on for a while then execute again. It will trigger a new batch job in EMR-S and very slow.

What is the expected behavior?
Increase the TTL if it has. Or identify why the batch job completed so soon.

What is your host/environment?

• OS: [e.g. iOS]
• Version [e.g. 22]
• Plugins

Do you have any screenshots?
If applicable, add screenshots to help explain your problem.

Do you have any additional context?
Add any other context about the problem.