Data Warehouse and Storage Import Monitoring and Error Guide

This guide explains every error that can appear on the Monitoring screen for data warehouses (Snowflake, BigQuery, Databricks) and Amazon S3 import integrations: what each error means, what happens to your data, why it happens, and how to fix it.

This guide covers imports only. Export integrations do not produce these errors.

How to read the monitoring screen

Each import run is summarized on the Monitoring screen:

• Success rate: The percentage of your data that was successfully ingested. It is calculated per cell (individual mapped values), not per row: successful cells ÷ total cells. A single bad column in a row does not fail the whole row's other columns.

• Error list: The errors found in the run, grouped by type, with a count for each. You can open a run to see the detailed breakdown.

• Columns that are never counted as errors: The cursor/timestamp column used to detect new data, and any column in your source that you did not map, are ignored. They are not shown as errors and do not lower your success rate.

Data errors

NoIdentifier

• What it means: The record has no identifier value, so we cannot tell which user it belongs to.

• What happens to your data: The whole record is skipped.

• Why it happens: The mapped identifier column is empty, null, or missing for that row.

• How to fix it: Make sure every row has a non-empty value in the column mapped as the identifier (for example email, phone, or your user ID).

InvalidIdentifier

• What it means: The identifier value is present but not in a valid format.

• What happens to your data: The whole record is skipped.

• Why it happens: The value does not match the expected format for its identifier type (for example an email that isn't a valid email address, or a malformed phone number).

• How to fix it: Clean the identifier values in your source so they match the expected format (valid email, correctly formatted phone number, etc.).

IdentifierConflict

• What it means: The identifiers in a single record point to more than one user, or conflict with each other.

• What happens to your data: The whole record is skipped.

• Why it happens: When more than one identifier is mapped, their values in the same row don't consistently describe one user.

• How to fix it: Ensure the identifier values in each row belong to the same person. When you map multiple identifiers, their combination must uniquely and consistently identify one user.

NoEventAndAttribute

• What it means: The record contains neither a valid event nor any attribute value — there is nothing to import.

• What happens to your data: The whole record is skipped.

• Why it happens: Every mapped payload field is empty for that row, or the mapping doesn't point at any populated column.

• How to fix it: Make sure each row carries at least one valid event or one attribute value, and that your mapping points to columns that actually contain data.

AttributePaused

• What it means: The attribute is currently paused in your integration settings, so its data is intentionally not imported.

• What happens to your data: That field is skipped for all records.

• Why it happens: The attribute was toggled off in the integration's attribute settings.

• How to fix it: If you want this data imported, un-pause the attribute in the integration's Attributes settings.

AttributeNotEligible

• What it means: The attribute is not eligible to be imported for this integration.

• What happens to your data: That field is skipped for all records.

• Why it happens: The attribute is not enabled/eligible in your account configuration for this import type.

• How to fix it: Confirm the attribute is eligible in your configuration, or map the field to an eligible attribute instead.

InvalidAttribute

• What it means: The attribute values fail validation.

• What happens to your data: That field is skipped for the affected records.

• Why it happens: The values don't conform to the attribute's rules (wrong type after normalization, or values that break validation).

• How to fix it: Correct the values in your source so they match the attribute's expected type and constraints.

NumberParsing

• What it means: A value mapped to a numeric attribute could not be read as a number.

• What happens to your data: That field is skipped for the affected records.

• Why it happens: The value contains characters that aren't valid in a number (letters, symbols, thousands separators, currency signs, blanks).

• How to fix it: Make sure values mapped to number attributes are clean numeric values.  

  • Snowflake / BigQuery / Databricks: cast or clean the column in your query (e.g. remove currency symbols/commas, TRY_CAST to a numeric type) so it returns plain numbers.

  • S3: ensure the values in the file are plain numbers (no currency or thousands separators).

TimestampParsing

• What it means: A value mapped to a timestamp attribute could not be read as a valid timestamp.

• What happens to your data: That field is skipped for the affected records.

• Why it happens: The value doesn't match the timestamp format the integration expects.

• How to fix it: Reformat the value to a recognized timestamp (ISO 8601, e.g. YYYY-MM-DD HH:MM:SS, or a Unix timestamp) that matches the format configured on the integration.  

  • Snowflake / BigQuery / Databricks: confirm the column is a genuine timestamp type, or format it to the expected string in your query.

  • S3: ensure the file's timestamp values use the configured format.

DateTimeParsing

• What it means: A value mapped to a datetime attribute could not be read as a valid date/time.

• What happens to your data: That field is skipped for the affected records.

• Why it happens: The date/time value doesn't match the expected format.

• How to fix it: Reformat the value to match the integration's configured date/time format. (Same platform guidance as TimestampParsing above.)

ValueTooLong

• What it means: A text value is longer than the maximum allowed length.

• What happens to your data: That field is skipped for the affected records.

• Why it happens: The string exceeds the attribute's length limit.

• How to fix it: Shorten or truncate the value in your source so it fits within the limit.

InvalidEventName

• What it means: The event name isn't a recognized event.

• What happens to your data: The affected event values are skipped.

• Why it happens: The event name has a typo or isn't defined in your event catalog.

• How to fix it: Check the event name in your source against your valid event names, and correct the typo or define the event. (The monitoring screen links to the relevant help article to find valid event names.)

InvalidEvent: EventParamInvalid

• What it means: An event parameter is missing or invalid.

• What happens to your data: The affected event is skipped.

• Why it happens: A parameter the event needs is absent or fails validation.

• How to fix it: Make sure all required event parameters are present and valid. The error detail names the parameter that caused the problem.

InvalidEvent: TimestampInvalid

• What it means: The event's timestamp could not be read.

• What happens to your data: The affected event is skipped.

• Why it happens: The event timestamp doesn't match the expected format.

• How to fix it: Reformat the event timestamp to match the integration's configured format (ISO 8601 or Unix timestamp).

FieldIsRequiredForEvent

• What it means: A required parameter for the event is missing.

• What happens to your data: The affected event is skipped.

• Why it happens: The mapping marks a parameter as required, but the source row doesn't provide it.

• How to fix it: Add the missing required parameter to your source data, or review whether it should be required in the mapping. The error detail names the missing parameter.

When an event has more than one invalid required parameter (for example both quantity and currency for purchase event), only one error is reported at a time; whichever parameter is checked first, which is not deterministic. After you fix that one, a different parameter error may appear on the next run. Review all of an event's parameters together rather than fixing them one at a time.

An event that is dropped because it falls outside the allowed time-to-live (TTL) window is not shown as a line error on the monitoring screen. If some events are missing but you don't see a matching error, check whether their timestamps are older than the allowed time window.

Array-of-object errors (Amazon S3 only)

KeyNotInMapping

• What it means: An object in an array-of-object field contains a key that isn't part of the attribute's mapping.

• What happens to your data: The unexpected key is skipped; the rest of the object is imported.

• Why it happens: The JSON objects in your array contain keys that aren't defined in the array-of-object attribute schema.

• How to fix it: Make the keys in your source objects match the attribute schema — remove unexpected keys, or add them to the mapping.

Failed syncs (Action required)

When a sync has a Failed status, it never started: it stopped on an integration-level problem — not a data issue — and imported nothing, so its Success Rate shows N/A. The integration moves to Action required and stays interrupted until you resolve the cause. Open the sync's Errors tab to see exactly which error it hit, fix it at the source platform (not in the data), then resume the sync.

The most common causes are:

• A permission was revoked — Insider can no longer read your source.  

  • Snowflake: Re-grant the role's privileges on the warehouse/database/schema.

  • BigQuery / GCS: Restore the service account's IAM roles in the Google Cloud Console.

  • Databricks: Restore the token/service-principal Unity Catalog grants.

  • S3: Restore the bucket policy / IAM permissions.

  • Azure Blob: Restore the connection string or RBAC role.

• Change tracking was disabled, or the stream became stale (Snowflake Streams / Databricks Change Data Feed): Re-create or re-enable change tracking so incremental syncs can resume.

• A mapped identifier column was dropped: Restore the column or update the mapping.

Quick reference