feat: add configurable timeout and expose metrics for routing v1

[SAHU-01]

Closes #115

Summary
Implements configurable timeout for HTTP routing v1 client to prevent denial of service when routing responses take longer than the previously hardcoded 15 seconds.

Note: HTTP routing metrics were already implemented and working in the existing codebase this PR only adds the missing configurable timeout functionality.

Changes Made
✅ Configurable HTTP Client Timeout

• Added RoutingV1HTTPClientTimeout config field
• Added --routing-v1-http-timeout CLI flag (follows naming from feat!: independent dht and routing v1 flags #113)
• Replaced missing HTTP client timeout (was infinite wait) with configurable value
• Replaced hardcoded 15s ParallelRouter timeout with configurable value
• Set default to 30s (safer than 15s for production, infra timeout ~1m)
• Added environment variable support: RAINBOW_ROUTING_V1_HTTP_TIMEOUT

✅ Verified Existing Metrics Work

• Confirmed existing httpclient.OpenCensusViews registration exposes metrics correctly
• Verified routing_http_client_latency and related metrics appear at /debug/metrics/prometheus
• Tested timeout configuration affects error metrics as expected

Testing

• CLI flag and environment variable work correctly
• HTTP client timeout prevents infinite waits
• ParallelRouter timeout configurable (was hardcoded 15s)
• Metrics show DeadlineExceeded errors with short timeouts
• Default timeout increased from 15s to 30s

Reviewer: @lidel and @hsanjuan

[hsanjuan]

Triage:

• @hsanjuan will review
• @hsanjuan should look if there is a timeout in kubo.

[hsanjuan]

This is setting a timeout for both the http.Client that makes requested to delegated routing endpoints, as well as the ParallelRouter.

Since we are introducing it, it's probably best to separate both timeout and have a RAINBOW_ROUTING_TIMEOUT next to RAINBOW_HTTP_ROUTERS_TIMEOUT.

[lidel]

Triage note:

• @hsanjuan will review once he has free time