feat: implement BullMQ Dead-Letter Queue (DLQ) and Admin Replay API #132

[Husten150]

Description

Context
The SubStream indexer currently lacks fault tolerance for malformed XDR data. A single corrupted ledger block can cause a worker hang, leading to data gaps for merchants. This PR introduces a Dead-Letter Queue (DLQ) to isolate "poison pill" payloads while allowing the ingestion pipeline to maintain 100% uptime.

Changes

Resilient Ingestion Flow: Wrapped the XDR parsing logic in a try/catch block with a 3-tier retry strategy.

DLQ Routing: Implemented automatic routing to a secondary failed-events-dlq after 3 failed attempts. Each entry includes the raw XDR payload, ledger sequence, and the error stack trace.

Non-Blocking Progression: Ensured the last_ingested_ledger pointer is updated even when a job is sent to the DLQ, preventing infinite loops on bad data.

Admin Replay API: Created a POST /admin/dlq/retry endpoint to allow manual re-injection of failed jobs into the primary queue.

Slack Integration: Added a Slack webhook notification that triggers specifically when a job is moved to the DLQ.

Data Retention: Configured BullMQ to retain DLQ messages for 14 days before automatic eviction.

[drips-wave]

Hey @Husten150! 👋 It looks like this PR isn't linked to any issue.

If this PR is for one of the issues assigned to you as part of a Wave, please link it to ensure your contribution is tracked properly. You can do this by adding a keyword to the PR description (e.g., Closes #123), or by clicking a button below:

Issue	Title
#132	Build Dead-Letter Queue (DLQ) for Failed RPC Syncs	Link to this issue
#131	Implement Idempotent Soroban Event Indexer Worker	Link to this issue
#134	Sync External Fiat-Price Oracle Cache (SEP-40)	Link to this issue
#133	Design PostgreSQL Schema for Subscription Cache	Link to this issue

ℹ️ Learn more about linking PRs to issues