Add provider address to event data structs

[AntiD2ta]

Changes

• Added Provider string field to all owned event data structs: HeadEvent, BlockEvent, ChainReorgEvent, FinalizedCheckpointEvent, BlobSidecarEvent, BlockGossipEvent, PayloadAttributesEvent, DataColumnSidecarEvent, and the generic Event container
• Added WithProvider/Provider context helpers in api/v1 for universal provider extraction across all event types
• HTTP and multi-client event handlers now populate the provider address on event data and context before forwarding to consumers

Note that seven event types use spec structs that model consensus data types (VersionedAttestation, AttesterSlashing, etc.) that should not carry a Provider field.

The context-based approach (apiv1.WithProvider/apiv1.Provider) provides a universal mechanism that works for all 15 event types, while the struct field is a convenience for the 9 types we own.

Example usage on Vouch

When the provider is available directly on the HeadEvent struct:

  func (s *Service) HandleHeadEvent(ctx context.Context, event *apiv1.HeadEvent) {
      // The provider address is embedded in the event data.
      provider := event.Provider // e.g. "http://localhost:5052"

      s.blockReceiptDelay.WithLabelValues(provider).Observe(delay.Seconds())
  }

For event types that use external spec structs (e.g. attestations), where the struct can't carry the field, we can extract it from context instead:

  func (s *Service) HandleAttestationEvent(ctx context.Context, event *spec.VersionedAttestation) {
      provider := apiv1.Provider(ctx) // universal mechanism, works for all event types

      // ...
  }

The context mechanism (apiv1.Provider(ctx)) works for all 15 event types, so Vouch could use it universally if preferred. The struct field is a convenience for the 8 owned types.

[Bez625]

I'm not a fan of the provider being passed in the context. I looked at vouch and I can only see us using head and block event handlers. I think let's limit this to updating the *Event structs to have the Provider field and set them where possible. Given we have explicit types at the callsite for these, we know if they have the Provider or not so it's type safe / requires no validation.

I think if we want this behaviour on the other handlers we should look to find reasons and upgrade them to returning *Event types that explicitly set the Provider.

[AntiD2ta]

Done 346dc29

[AntiD2ta]

Also: #286 (comment)

[AntiD2ta]

Created #288 to address adding the Provider on consensus types event data structs.