Documentation for accessing Fivetran sync logs via your data warehouse.

Fivetran does not provide a REST API for accessing sync logs. Instead, sync logs and operational metadata are available through the Fivetran Platform Connector, which syncs this data directly to your data warehouse.

Overview

To access sync logs, you need to set up the Fivetran Platform Connector:

1. Create a new connector in your Fivetran dashboard
2. Select "Fivetran Platform" as the source type
3. Configure it to sync to your destination warehouse
4. The connector will populate metadata tables with sync history

Schema Location

By default, Fivetran Platform Connector data is stored in a schema named fivetran_metadata (or fivetran_log for legacy setups). The schema name is configurable when setting up the connector.

Key Tables

The Fivetran Platform Connector provides several tables:

LOG

The primary table for sync event logs. Contains:

• id - Unique log entry identifier
• connector_id - The connector that generated the log
• event - Event type (e.g., "sync_start", "sync_end", "write_to_table")
• message_event - Detailed event category
• message_data - JSON with event-specific details
• time_stamp - When the event occurred

CONNECTOR_SDK_LOG

Logs from SDK-based connectors (custom connectors). Contains:

• connector_id - The SDK connector ID
• log_message - The log message text
• log_level - Log severity (INFO, WARNING, ERROR)
• time_stamp - When the log was generated

AUDIT_TRAIL

Account-level audit events. Contains:

• id - Unique audit entry identifier
• actor - Who performed the action
• action - What action was performed
• object_type - Type of object affected
• object_id - ID of the affected object
• time_stamp - When the action occurred

Example SQL Queries

Recent Sync Events

Get the last 10 sync events for a specific connector:

SELECT
  time_stamp,
  event,
  message_event,
  message_data
FROM fivetran_metadata.log
WHERE connector_id = 'your_connector_id'
ORDER BY time_stamp DESC
LIMIT 10;

Failed Syncs

Find recent failed syncs:

SELECT
  connector_id,
  time_stamp,
  message_event,
  message_data
FROM fivetran_metadata.log
WHERE event = 'SEVERE'
  OR message_event LIKE '%error%'
  OR message_event LIKE '%fail%'
ORDER BY time_stamp DESC
LIMIT 50;

Sync Duration

Calculate sync duration for recent syncs:

WITH sync_events AS (
  SELECT
    connector_id,
    time_stamp,
    message_event,
    LAG(time_stamp) OVER (
      PARTITION BY connector_id
      ORDER BY time_stamp
    ) AS prev_time
  FROM fivetran_metadata.log
  WHERE message_event IN ('sync_start', 'sync_end')
)
SELECT
  connector_id,
  time_stamp AS sync_end_time,
  DATEDIFF('minute', prev_time, time_stamp) AS duration_minutes
FROM sync_events
WHERE message_event = 'sync_end'
  AND connector_id = 'your_connector_id'
ORDER BY time_stamp DESC
LIMIT 20;

Schema Changes

Find schema modification events:

SELECT
  connector_id,
  time_stamp,
  message_event,
  message_data
FROM fivetran_metadata.log
WHERE message_event LIKE '%schema%'
  OR message_event LIKE '%column%'
  OR message_event LIKE '%table%'
ORDER BY time_stamp DESC
LIMIT 100;

Usage with Fivetrex

While Fivetrex cannot query sync logs directly (no REST API exists), you can use the connector state and status information:

# Get current sync state
{:ok, connector} = Fivetrex.Connectors.get(client, "connector_id")
Fivetrex.Models.Connector.sync_state(connector)

# Get sync status summary
{:ok, status} = Fivetrex.Connectors.get_sync_status(client, "connector_id")

For historical sync logs and detailed event data, query the fivetran_metadata schema in your data warehouse using the SQL examples above.

See Also

• Fivetrex.Connectors - For real-time connector status
• Fivetrex.Models.Connector - Connector struct with status helpers
• Fivetran Platform Connector docs