Kubernetes Logging with Fluentd/Fluent Bit

Help me set up centralized logging for my Kubernetes cluster.

Cluster details:
- Cluster size: [nodes, pods]
- Log volume: [estimated GB/day]
- Log destinations: [Elasticsearch/CloudWatch/Stackdriver/Loki]
- Retention requirements: [days]

Logging requirements:
1. Log collection
   - Fluent Bit DaemonSet
   - Container log parsing
   - Kubernetes metadata enrichment
   - Custom log formats

2. Log processing
   - Filtering and transformation
   - Multi-line log handling
   - JSON parsing
   - Field extraction

3. Log routing
   - Multiple outputs
   - Conditional routing
   - Buffering and retry
   - Dead letter queue

4. Operations
   - Resource limits
   - Backpressure handling
   - Health monitoring
   - Performance tuning

Include:
- Fluent Bit ConfigMap
- DaemonSet manifest
- Parser configurations
- Destination-specific output configs