Add metrics for tenant attribution and write requests

[willh-db]

Changes

Three new Prometheus metrics to support the migration from m3coordinator to Thanos-native tenant attribution.

1. thanos_receive_write_requests_total
   • Tracks all write requests.
   • Labels: code (HTTP status), tenant (from HTTP header)
   • Replaces coordinator_write_success / coordinator_write_errors for SLO tracking
2. thanos_receive_write_rejected_total
   • Tracks rejected write requests by reason.
   • Labels: reason (write_gate, head_series_limit, request_size, series_limit, samples_limit), tenant (from HTTP header)
   • Replaces coordinator_write_reject for rejection monitoring
3. thanos_receive_tenant_attributed_total
   • Tracks per-timeseries tenant attribution.
   • Labels: tenant (attributed), source (http_header/label_rules)
   • Provides visibility into how/which tenants are being attributed, whether from explicit HTTP headers or rule-based label matching

Verification