Troubleshooting: Ingestion Task Failed

Issue

An ingestion task has failed, and the dataset has not been updated as expected. The ingestion status shows FAILED on the Ingestion Tasks view.

Understanding Ingestion Failures

When an ingestion task fails, the Connector first attempts to retry automatically. While retrying, the task status shows RETRYING — this indicates automatic retry is in progress and no action is needed. Wait until the task either recovers (transitions to SUCCESS) or exhausts all retry attempts. Once all retries are exhausted, the Connector marks the task as FAILED and automatically disables it to prevent repeated unsuccessful attempts. Before taking action on a FAILED task, it helps to understand whether the failure is transient (likely to succeed on retry) or persistent (requires a configuration change or is a known limitation).

FAILED at Creation vs. FAILED After Retries

An ingestion task can show FAILED in two different situations, distinguishable by its run history in the Task Details view:

• FAILED at creation (empty run history): the ingestion task itself could not be fully set up, and no ingestion run was ever attempted. This happens when the Connector could not reach the data service while creating the task, or when a plan limit (such as the Free edition's ingestion task limit) was reached. When you try to enable ingestion, an alert message explains what happened, and hovering over the information icon next to the task's FAILED status shows the same reason. Once the underlying issue is resolved, re-enable the task to try again.
• FAILED after retry exhaustion (one or more runs in history): the task attempted ingestion — possibly more than once — and failed after exhausting its automatic retries. The rest of this article covers this case in detail.

RETRYING Status Display

When a task shows RETRYING, two places in the Connector surface retry detail. There are two kinds of retry:

• Retrying shortly — the Connector is trying again within a few minutes. No attempt count is shown.
• Retry 1/2, Retry 2/2 — the Connector is trying again on a following day, and the count shows the attempt out of the maximum allowed (default 2).

Depending on the failure, a task may go through the first kind and then the second, or only the second. An on-demand task is marked FAILED after any retries within minutes.

Ingestion Tasks list — info tooltip (orange icon next to RETRYING):

Click or hover the information icon to see a summary. For a retry within minutes:

Retrying shortly — timeout connecting to data source

For a retry on a following day:

Retry 1/2 — next attempt 2026-04-08 06:00:00 — timeout connecting to data source

Part	Meaning
Retrying shortly	The next attempt happens within a few minutes. Shown instead of a retry count.
Retry 1/2	Current retry attempt out of the maximum allowed (default 2), for retries on following days.
next attempt <time>	The scheduled time of the next retry. Shown only when a next retry time is available.
error message	The reason the previous attempt failed.

Task Details view — "Next Retry" card:

Open the task in the Task Details view to see the "Next Retry" card, which shows the scheduled time of the next retry attempt alongside "Retrying shortly" or the retry count, last error message, and the full ingestion run history. This card replaces the standard "Next Ingestion" card while the task is in RETRYING status.

Delete says the task is still finishing its current run

If deleting an ingestion task answers that the task is still finishing its current run, nothing was deleted and the task is still listed. Wait a moment and delete it again.

Transient Failures

These failures are caused by temporary conditions and typically resolve on their own. Re-enabling the task is usually sufficient.

• Data source temporarily unavailable — The geospatial service is down for maintenance or experiencing an outage. This is the most common failure cause.
• Timeout during data retrieval — The data source responded too slowly. Large or complex datasets may occasionally exceed processing time limits, but succeed on a subsequent attempt under lighter server load.
• Network connectivity issues — A temporary network disruption between the Connector and the data source.

Persistent Failures

These failures will recur on every retry unless the underlying cause is addressed.

• Non-standard service implementation — The data source lists datasets that cannot actually be retrieved. The server may reject the data request, require non-standard query parameters, or have a server-side misconfiguration.
• Empty service — The service endpoint responds successfully but lists no datasets.
• Unsupported file content — The data source provides a compressed file (ZIP, GZIP) that does not contain any supported geospatial formats. The Connector cannot process the contents and marks the task as a permanent failure. See Unsupported File Content below.
• Dataset exceeds current size limits — Raster coverages (WCS) that exceed the current maximum ingestion size limit cannot be downloaded in a single request. See WCS Coverage Size Limits below.
• Invalid or corrupted source data — The source dataset contains geometry or attribute data that cannot be parsed or converted.
• Permissions or authentication errors — The data source requires credentials that are missing, expired, or invalid. Check the connection's authentication settings in the Connections view (see Managing Connections).
• Application permissions revoked — One or more Geo Data Connector application permissions granted at installation have been revoked within Snowflake.

WCS Coverage Size Limits

WCS (Web Coverage Service) coverages currently have a maximum ingestion size limit. Coverages that exceed this limit cannot be ingested and will fail immediately with no data loaded. This limit is expected to be raised in a future release.

Retrying the task will produce the same failure. If you encounter this issue, consider whether a smaller geographic subset of the coverage would meet your needs. Contact support for assistance with geographic subsetting to ingest a specific region of the coverage instead of the full extent.

Note: This limit applies specifically to WCS coverages. WFS, WMS, and WMTS data sources handle large datasets through pagination and tiling, so they are not subject to the same constraint.

Unsupported File Content

Some data sources provide downloadable files as compressed archives (ZIP, GZIP). If the Connector downloads a compressed file and finds that it does not contain any recognized geospatial data formats, the task fails permanently with the message:

"This compressed file does not contain any supported geospatial formats."

This failure is non-retryable — re-enabling the task will produce the same result. The compressed archive from the data source does not include files in formats that the Connector can process (such as GeoJSON, GML, Shapefile, or CSV with geographic data).

If you encounter this issue, verify the data source to confirm what file formats it provides. Consider whether an alternative dataset or endpoint from the same provider offers the data in a supported format.

Diagnosing Failures

Check the Task Details View

Open the failed task in the Task Details view to review the ingestion run history. Key indicators:

Indicator	What It Tells You
Single failed run, no prior successes	The dataset may have a persistent issue (size limit, incompatible format, or unreachable data).
Failed after previous successes	Likely a transient issue — the data source was reachable before but failed this time. Re-enabling usually resolves it.
Multiple consecutive failures	Suggests a persistent problem. Check if the data source has changed or if a new limitation applies.
Rows Loaded = 0 on failed run	The failure occurred before any data could be retrieved — typically a connectivity, permissions, or size limit issue.
Rows Loaded > 0 on failed run	Partial data was retrieved before failure — may indicate a timeout or data quality issue mid-transfer.

Verify the Data Source

Test the data source URL in the Explorer view or directly in a browser:

• If the service does not respond, the failure is transient — wait and retry.
• If the service responds but the specific dataset is no longer listed, the data source may have been reconfigured.
• If the service responds normally, the issue may be specific to the dataset format, size, or a non-standard server behavior.

Resolution Steps

1. Retry the Task (Transient Failures)

If the failure appears transient, manually re-enable the ingestion task by toggling it OFF and then ON again in the Tasks view or Dataset view. The task will run again at its next scheduled time, or you can trigger it immediately using the action icon.

If the data service is still unreachable at the moment you retry, re-enabling will not succeed — the task remains FAILED, and the information icon next to its status keeps showing the reason so you know a retry is still needed rather than assuming it worked. Wait a little longer and try again once the data source is reachable.

2. Verify Data Source Availability

Confirm that the geospatial data source is reachable and responsive by testing the URL in the Explorer view or directly via a browser or API client.

3. Check Network and Permissions

Ensure that network settings allow Snowflake to access the data source and that any required authentication credentials or tokens are valid and up to date.

4. Verify Application Permissions

Contact your Snowflake system administrator to verify that all necessary Geo Data Connector application permissions are intact. Permissions can be reviewed in Snowsight via the main menu option Catalog, submenu Apps.

Compare the current permissions with those required by Geo Data Connector as outlined in the Installation chapter.

Tip: When reviewing permissions, consider granting the optional permissions that allow the Connector to share logs with the application provider, as this can assist in troubleshooting and support.

A manual run could not start

If you select Run now and the Connector cannot complete the start, the task can show as suspended, with the schedule switch off and Resume available. That is the task's true state: the Connector turned the schedule off rather than leave a task that looks active but never runs. Select Resume to turn the schedule back on.

Running a task that you suspended yourself runs it once and leaves it suspended; use Resume when you want its schedule to run again. If that single run fails, the Connector does not retry it automatically: the task shows FAILED with the reason, and you can run it again when you are ready. (A task that runs on demand is still retried within minutes, as before.)

A task was switched off because it is no longer set up

If a scheduled run finds that the task is no longer set up, the Connector does not start an ingestion run. It switches the task off and shows the status INACTIVE with an information icon; hover over it to read the reason: "This task is no longer set up, so no ingestion run was started and the task is disabled. Resume it to try setting it up again; if that does not work, delete the task and create it again." If notifications are enabled for task failures, you also receive an e-mail with the same advice.

Select Resume to try setting the task up again. If the task still cannot be resumed, delete it and create it again; data already ingested is kept.

A task that you suspended yourself never shows this reason, even if it failed earlier.

5. Contact Support

If the issue persists after performing the recommended checks, please contact the Geo Data Connector support team. The support email address is displayed on the application page in Snowflake Marketplace within Snowsight.

When reaching out, please provide detailed information about:

• The data source URL and dataset name
• The data source type (WFS, WMS, WMTS, or WCS)
• The ingestion task details — including the task name, status, and timestamps from the Task Details view
• Whether the task has previously succeeded or has always failed

Including these details will help the support team diagnose and resolve the issue more efficiently.

Discovery Errors

Discovery failures occur when the Connector cannot explore a data source. These are separate from ingestion task failures — discovery fails before a task is created.

Authentication Failed

"Authentication failed. Verify that the connection has the correct credentials configured."

This error appears when the data source requires authentication but the credentials are missing or incorrect. To resolve:

1. Open the Connections page
2. Find the connection for the affected service
3. Verify the authentication settings (username and password)
4. Re-submit the discovery request

If the data source does not have a saved connection, create one first with the correct authentication settings. See Managing Connections for details.

Connection or Timeout Errors

"Could not connect to the service" or "The service did not respond in time"

These errors indicate that the data source is unreachable or too slow to respond. Verify that the URL is correct and the service is available by testing it in a browser. If the service is temporarily down, try again later.