Graceful shutdown support

[lfrancke]

As a user, I want to minimize disruptions of my deployed products.
There are important items listed in https://docs.stackable.tech/home/nightly/concepts/operations/#_service_availability, but for one very important we are missing the actual implementation in all operators but trino: Graceful shutdown

We should check all our products on whether there is anything we can do to shutdown gracefully once a Pod has been scheduled for deletion.

Tasks

Preview Give feedback

1. Support graceful shutdown airflow-operator#343
   8 of 12
   release/2023-11 +1
   
2. Support graceful shutdown druid-operator#486
   8 of 12
   release/2023-11 +1
   
3. Support graceful shutdown hbase-operator#400
   release/2023-11 +1
   
4. feat: Support graceful shutdown hdfs-operator#407
   10 of 12
   +0
   
5. feat: Support graceful shutdown hive-operator#385
   10 of 12
   +0
   
6. feat: Support graceful shutdown kafka-operator#635
   11 of 12
   release/2023-11 +1
   
7. Support graceful shutdown nifi-operator#528
   9 of 12
   release/2023-11 +1
   
8. feat: Support graceful shutdown opa-operator#487
   6 of 12
   release/2023-11 +1
   
9. docs: Document graceful shutdown spark-k8s-operator#306
   1 of 12
   release/2023-11 +1
   
10. Support graceful shutdown superset-operator#422
    8 of 12
    release/2023-11 +1
    
11. refactor!: Move gracefulShutdownTimeout to roleGroup config trino-operator#486
    5 of 12
    +0
    
12. feat: Support graceful shutdown zookeeper-operator#740
    11 of 12
    release/2023-11 +1
    
13. Support graceful shutdown hello-world-operator#33
    8 of 12
    +0
    
14. feat: Document and clarify vector graceful shutdown operator-rs#681
    2 of 8
    +0
    
15. docs: Fix PID 1 description hdfs-operator#418
    0 of 12
    +0
    
16. docs: Fix PID 1 description hbase-operator#410
    0 of 12
    +0
    

Tasks

Preview Give feedback

1. Approve CRD changes
2. feat: impl Atomic for Duration operator-rs#668
   0 of 10
   +0
   
3. Add concepts guide on graceful shutdown documentation#468
   release/2023-11 +1
   
4. refactor!: Change PodBuilder::termination_grace_period to take Duration struct operator-rs#672
   0 of 10
   +0
   

[soenkeliebau]

We'll put this on the board to do some research into it.

HBase and HDFS are the main two products we want this for, after those are done we'll revisit and see if we go on with the epic or put it back into the backlog again.