spec-msg/testing: End to end testing and rollout readiness

[lexnv]

Overarching issue for everything that gates the rollout.

This includes unit-testing, zombinet end to end testing and observability tooling.

1. spec-msg/tests: End to end testing for HRMP replacement #12596

• Add 2 penpal runtimes over spec-msg with full flow
• First use HRMP as baseline, then cutover and drain over to SpecMsg, then rollback
• Include KPI, poc: https://github.com/paritytech/polkadot-sdk/blob/64dc8878a313ee960a4067adf35e5db8dec005a6/polkadot/zombienet-sdk-tests/tests/messaging/spec_msg_penpal.rs

2. Fuzzing hardening into CI

• Fuzz target for consensus decoders (stream ID, lifs, wire messages, V3 block data)
• e2e suite of tests:
  • receiver downtime
  • lift catchup
  • resubmission
  • reorg storms
  • elastic scaling bursts
  • restart persistance

3. spec-msg/monitoring: Add monitoring metrics for operational readiness #12597

   • Monitoring and tooling for readiness
   • Add spec_msg_* class of prometheus metrics
   • Subxt on chain explorer for:
     • ring occupancy
     • channel phases
     • register watermarks and credits
   • Canary testing for live HRMP <-> Spec Msg cycles
   • Grafana dashboards and alert rules

4. Security Audit

   • Scope: pallets, relay spec-msg module, validate_block wrapper, lift verifications
   • Happens as soon as the runtime surface freezes