apollo_deployments: load each service's app configs through the node entry point

[matanl-starkware]

Addresses yoavGrs' line-132 comment on #15083. Stacked on it — rebase onto main-v0.14.4 once that merges.

Making the merge helper pub is the wrong route: update_config_map type-checks, and 116 app-config values are $$$_…_$$$ placeholders that fail it. Stub those first and the whole production entry point works instead — SequencerNodeConfig::load_and_process, the same call a pod makes, no visibility changes.

Catches stale params, wrong types and pruning errors.

[cursor]

PR Summary

Low Risk
Test-only changes plus a small refactor to share replacer placeholder formatting; no production deployment or runtime behavior changes.

Overview
Adds all_service_configs_load, which exercises every node service’s merged application configs through SequencerNodeConfig::load_and_process—the same path pods use at startup—plus a default secrets file.

Before loading, deploy_time_value substitutes environment-specific values: $$$_…_$$$ replacer placeholders become schema defaults, and PointerTarget entries become small well-formed stubs (e.g. SN_MAIN, localhost URLs, 0x1 addresses).

replacers.rs exposes the placeholder string logic as replacer_annotation so tests and replacer generation share one definition instead of a private format_key.

^{Reviewed by Cursor Bugbot for commit c2665f3. Bugbot is set up for automated code reviews on this repo. Configure here.}

[reviewable-StarkWare]

This change is