Request data older than the retention_period result as an error

[eraac]

Describe the bug
If via Grafana you request 30 days of logs with a retention_period of 2 weeks, Loki return an error

To Reproduce
Steps to reproduce the behavior:

1. Started Loki (2 weeks of retention_period for the index, store in dynamoDB)
2. Query last 30 days of data

Expected behavior
Loki shouldn't try to read deleted index (and generate error)

Environment:

• Infrastructure: Kubernetes
• Deployment tool: helm
• Loki version: 1.2

Logs in Grafana

QueryPage error: table=loki_index_2605, err=ResourceNotFoundException: Requested resource not found status code: 400, request id: SJQA2EBBC38QBE9G0QT2FBH8PNVV4KQNSO5AEMVJF66Q9ASUAAJG

loki_index_2605 is an old table (dynamoDB), ~ 5 weeks old I guess, so with a retention_period of 2 weeks this table doesn't exist anymore

[sandeepsukhani]

While updating retention, ideally you should also change max look back period here.
Do you want to contribute by adding a check for it and update relevant documentation?