Repair Search Index Drift

The problem

Search results frequently fall out of sync with your database due to lost queue events, race conditions, or failed reindexing. Users see "ghost" results that were deleted or miss new records entirely because indexing updates were reordered or dropped.

What it does

• Detects missing, stale, and duplicate search documents by reconciling index state against your authoritative database.
• Implements version-aware indexing to prevent older updates from overwriting newer data.
• Automates safe, zero-downtime index rebuilds using alias-based cutovers and validation checks.
• Repairs broken deletion logic by implementing stable tombstones and idempotent event processing.

Frameworks & tools

Elasticsearch, OpenSearch, Algolia, Meilisearch, and Typesense. Works with SQL and NoSQL databases using queues, transactional outboxes, or CDC pipelines.

Why this beats prompting it yourself

Generic prompts often suggest a full "delete and reindex," which causes downtime and ignores the root cause of drift. This skill enforces a versioned consistency model that handles specific edge cases like reordered events and partial bulk-update failures that basic prompts miss.

Use cases

• Fixing search results that stay stale after a database update or restore.
• Removing deleted records that still appear in search results.
• Implementing a transactional outbox to ensure every database change hits your index.
• Migrating search mappings without losing data or risking downtime.

Known limitations

Requires an existing project workspace and access to both the database schema and search provider configuration. Cannot guarantee consistency if the database lacks a monotonic version or timestamp.