Skip to main content

​
Endpoint

POST /search/reindex
Triggers a full reindex operation that syncs all data from the PostgreSQL database to Typesense search collections. This process runs asynchronously to avoid HTTP timeouts.

​
Request Body

​
batch_size
integer
default:"1000"
Number of records to process in each batch. Larger values are faster but use more memory

​
Response

​
message
string
Confirmation message that the reindex operation has started
​
progress
object
Initial progress information

​
Example Request

curl -X POST "https://api.blnk.io/search/reindex" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"batch_size": 1000}'

{
  "batch_size": 1000
}

​
Example Response

{
  "message": "Reindex operation started",
  "progress": {
    "status": "in_progress",
    "total_records": 0,
    "processed_records": 0,
    "current_collection": "ledgers",
    "started_at": "2024-03-04T12:00:00Z"
  }
}

​
Monitoring Progress

Use the Get Reindex Progress endpoint to monitor the reindex operation: 
curl -X GET "https://api.blnk.io/search/reindex" \
  -H "Authorization: Bearer YOUR_API_KEY"

​
When to Reindex

You should trigger a reindex in the following situations:

​
Initial Setup

When first deploying Blnk with Typesense search enabled.

​
After Data Migration

After importing large amounts of data directly into the database.

​
Schema Changes

After updating search collection schemas or adding new searchable fields.

​
Data Inconsistencies

If you notice search results don’t match database records.

​
Recovery

After recovering from a Typesense outage or data corruption.

​
Reindex Process

The reindex operation processes collections in the following order:
  1. Ledgers: Base collection, reindexed first
  2. Identities: User and entity records
  3. Balances: Account balance records
  4. Transactions: Transaction history
  5. Reconciliations: Reconciliation records

​
Error Responses

​
error
string
Error message describing what went wrong
​
progress
object
Current progress information (returned when a reindex is already running)

​
Common Errors

409 Conflict: A reindex operation is already in progress 
{
  "error": "A reindex operation is already in progress",
  "progress": {
    "status": "in_progress",
    "total_records": 50000,
    "processed_records": 12500,
    "current_collection": "transactions",
    "started_at": "2024-03-04T11:45:00Z"
  }
}

​
Performance Considerations

​
Batch Size

  • Small (100-500): Lower memory usage, slower overall
  • Medium (500-1000): Balanced performance (recommended)
  • Large (1000-5000): Faster but higher memory usage

​
Database Impact

Reindexing reads from your database:
  • Uses read replicas if available
  • Minimal impact on write operations
  • May increase database CPU during reindex

​
Duration Estimates

RecordsBatch SizeEstimated Time
10,0001,0001-2 minutes
100,0001,00010-15 minutes
1,000,0001,0001-2 hours
10,000,0002,0008-12 hours

​
Best Practices

  1. Schedule during low traffic: Run reindex during off-peak hours
  2. Monitor progress: Check the progress endpoint regularly
  3. Use appropriate batch size: Balance speed and resource usage
  4. Don’t interrupt: Let the reindex complete naturally
  5. Have a backup: Ensure your database is backed up before major reindex operations

​
Incremental Updates

For normal operations, you don’t need to reindex. Blnk automatically updates the search index when:
  • New records are created
  • Existing records are updated
  • Records are deleted
Only use full reindex for the scenarios listed above.

Build docs developers (and LLMs) love

Get started for freeTalk to us