Fix Mammoth connection and data refresh failures

Read a connection's health status

Each connection has a health status on the Connections page. Health comes from a daily check and from real reads and exports. See Manage saved connections.

• Authentication failure: if a credential is revoked, rotated at the source or expires, the next refresh for that connection fails and the connection shows Error. The Connectors launcher lists it under Needs attention with the reason key expired.
• Other failures: source downtime, network issues and rate limiting also fail the run. An unreachable source shows unreachable under Needs attention. There is no automatic retry.

Fix an expired key or failed sign-in on a source connection

Import and export use separate credentials, so you fix them separately. When an import connection (PostgreSQL, MSSQL and other source connectors) loses authentication, the dataset refresh fails with an authentication error. Causes include an expired token, a rotated password or a revoked credential.

1. Open the failed automation and read Last run result for the reason.
2. Open the connection from Needs attention, or from the Connections page.
3. Click Edit connection and enter a new key or sign in again. Only a Project Admin can do this.
4. Run the automation again with Run now to confirm it works.

See Manage automations for what an Error status means on an automation.

Fix export credentials for a destination

An export fails when the write credentials for a destination (PostgreSQL, BigQuery and other export destinations) have expired or been revoked. The destination has its own credentials, separate from any source connection.

1. Open the export configuration and update the destination credentials.
2. Click Validate Connection in the export form to confirm the credentials work.
3. Apply the change and re-run the export.

If you send to a database on the Mammoth Managed tab, Mammoth manages its credentials. Contact support if those destinations fail. See Send to destinations.

Fix pipeline errors after a source change

When a source's columns change, tasks in the pipeline can fail. See Fix a failing step for each fix.

A slow run isn't an error: the pipeline still completes. It usually means the source grew a lot, or a task processes more rows than you expected. Find the slow task and add filters to cut the rows it handles. See Speed up pipelines and connectors.

Fix data that doesn't match the dataset's columns

When you add data to a dataset (through a Dataset Refresh, a Data Consolidation or a manual addition), Mammoth compares the incoming data with the existing columns. A missing, renamed or retyped column is handled by the dataset's If column type doesn't match and column mapping settings. Mammoth can convert the column to text, convert incompatible values to null, or leave the dataset unchanged. See Add or replace data.

A failed data pull shows in the notifications bell with the dataset's name. See Notifications.

To fix a failed addition, make the incoming data match the dataset's columns at the source, or adjust the dataset's mapping settings. Then retry.

Fix Send to Dataset column errors

Send to Dataset sends processed data from one View to a destination dataset. Its column mapping must match the destination dataset's columns. If a mapped column no longer exists in the source View or the destination dataset, the send fails.

1. Open the Send to Dataset configuration.
2. Confirm every mapped column exists in both the source View and the destination dataset.
3. Update the mapping and retry.

What happens when a refresh automation fails

When a Dataset Refresh automation fails because of a connection issue, it shows Error, and Last run result in its detail panel gives the reason.

• Mammoth does not retry a failed run. Fix the cause, such as an expired credential, then click Run now instead of waiting.
• A recurring schedule still runs at its next time, whatever happened before. That is the schedule continuing, not a retry.
• When one automation refreshes several datasets, Mammoth checks each connection separately. Some datasets can refresh while others fail.

Fix network and firewall problems

If a connection fails now and then and its credentials are valid, the cause may be the network.

1. Confirm your database or service accepts connections on its port.
2. Allow the Mammoth IP address shown in the connection form in your firewall.
3. Check the port. The defaults are 5432 for PostgreSQL, 3306 for MySQL and 1433 for SQL Server, and you can use a custom port.

A PostgreSQL server reachable only through a bastion host can connect through an SSH tunnel:

1. On the PostgreSQL form, open Connection options.
2. Tick Connect through an SSH tunnel.
3. Fill in SSH host, SSH port, SSH user and SSH sign-in.

See Allow Mammoth through your firewall for the IP address, ports and firewall steps for each connector.

Fix refreshes that fail on a source's rate limit

If a source limits how often it accepts requests, refreshes may fail until the limit resets. There is no automatic retry, the same as any other failure. The automation runs again at its next scheduled time, or you can click Run now once the limit resets.

To hit rate limits less often:

• Spread automation schedules so many refreshes don't hit the same source at the same time.
• Consolidate datasets where you can.

Where connection failures don't show up

The Activity Log records changes, such as connections being created, edited or deleted. It does not record a connection's health, so the failures on this page don't appear there. Check the automation's status, Last run result and the error shown in the interface instead. See Activity log.

Webhooks push data into Mammoth instead of Mammoth pulling it, so they aren't covered here. See Receive data by webhook.

When to contact support about a connection

Contact support@mammoth.io when errors persist after these steps, or when you need help with IP address allow rules or SSH tunnel setup. Include:

• the connector type
• the exact error message from the notification or error indicator
• a summary of the steps you've already tried

Related

• Manage saved connections: health status, editing and deleting connections.
• Refresh a dataset: set up the automation that pulls new data.
• Speed up pipelines and connectors: fix refreshes that time out or get slower each week.
• Error messages: the exact messages for automations, schema errors and permissions.
• Allow Mammoth through your firewall: IP addresses and ports by connector.