feat: enable multiple layers of caching for impersonation token.

Type of change

• Bug Fix
• Feature
• Pipeline
• Test
• Documentation

Does this introduce a change in the core logic?

• No
• Yes

Does this introduce a change in the cloud provider implementation, if so which cloud?

• AWS
• Anthos
• Azure
• GCP
• IBM

Does this follow conventional commits spec?

• No
• Yes

Have you set the target Milestone?

• No
• Yes

Have you set the no-detached-pipeline label?

• No
• Yes

Updates description?

enable multiple layers of caching for impersonation token.

• in-memory cache for caching the token in pod memory.
• shared redis cache when the cache is not available in pod memory.