Skip to content

Auto Manage Indexes

POST
/api/v1/indexes/auto-manage

Automatically manage indexes based on recommendations.

Analyzes all models and optionally applies recommendations:

  • Creates IVFFLAT indexes for models above threshold
  • Upgrades IVFFLAT to HNSW for models with 100K+ vectors
  • Rebuilds IVFFLAT indexes with suboptimal parameters

Args: request: Auto-manage request with dry_run flag service: Index management service (injected)

Returns: Results of auto-management with applied changes

Raises: HTTPException: 500 for server errors

AutoManageRequestModel

API request model for automatic index management.

object
dry_run
Dry Run

If true, only return recommendations without applying them

boolean

Successful Response

AutoManageResponseModel

API response model for automatic index management.

object
results
required
Results
Array<object>
AutoManageResultModel

API model for a single auto-management result.

object
model_name
required
Model Name
string
action
required
Action
string
reason
required
Reason
string
applied
required
Applied
boolean
index_info
Any of:
IndexInfoModel

API model for vector index information.

object
index_name
required
Index Name
string
model_name
required
Model Name
string
index_type
required
Index Type
string
Allowed values: ivfflat hnsw
row_count
required
Row Count
integer
size_bytes
required
Size Bytes
integer
size_pretty
required
Size Pretty
string
parameters
required
Parameters
object
key
additional properties
integer
scan_count
Scan Count
integer
0
total_count
required
Total Count
integer
applied_count
required
Applied Count
integer
dry_run
required
Dry Run
boolean

Validation Error

HTTPValidationError
object
detail
Detail
Array<object>
ValidationError
object
loc
required
Location
Array
msg
required
Message
string
type
required
Error Type
string
input
Input
ctx
Context
object