Skip to main content

Connection health checks

Guides

A health check runs a saved connection's test on a schedule. Each run tests every environment the connection has settings for, so a password that expires in Production is caught before a data pipeline needs it.

A failed health check does not stop or delay runs.

Set a schedule​

  1. Open the connection
    On the Connections page, select the connection's name, or choose Configure in its row menu.
  2. Find Health check
    The section sits below the connection's settings. It appears once the connection is saved.
  3. Choose how often
    Off, Daily or Weekly. Weekly also asks for a day.
  4. Set the time
    Times use your account's time zone, shown beside the time.
  5. Select Save schedule
    Next check shows when the first scheduled test starts.
  • Anyone who can edit the connection can change its schedule. Anyone who can see the connection can see its schedule and results.
  • Some connectors have no connection test. For those, the section reads Scheduled checks are not available for this connector.
  • Choose Off to stop scheduled tests. Past results stay.
  • An Inactive connection is not tested. Its schedule resumes when the connection is active again.

What each status means​

StatusMeaning
PassedThe test reached the system and the credentials were accepted.
FailedThe test ran and the system refused it, for example a wrong password or a missing permission. The message says why.
UnreachableNo answer within 60 seconds, or the Private Agent did not pick the test up within 10 minutes.

Where you see results​

  • Health check section: the newest result for each environment, with the message for a result that did not pass.
  • Connections list: each environment's chip on the row shows its newest result. Hover over a chip to see when it was tested. An environment that has never been tested shows a dashed circle.
  • History: choose Test history in the row menu, View history under a failed test's Details, or Show history in the Health check section, to see the last three results for each environment.
  • A test of a connection's base settings, with no environment, is listed as Base.

Each result shows:

  • The status and how long ago the test ran. Hover over the time to see the date and time in your account's time zone.
  • Scheduled or Manual. A manual result names the person who ran it.
  • Where it ran: Cloud or Private Agent.
  • The message, for a result that did not pass.

Selecting Test connection in the editor, or Test on a row, also records a result. Older results beyond three per environment are removed.

Connections reached through a Private Agent​

When your runs for an environment go through a Private Agent, that environment's health check runs on the agent too. The test takes the same network path your runs take, and its results read Private Agent.

Connections in an environment that runs on your Private Agent are tested from your agent, both on a schedule and from Test on the Connections page.

If your agent does not pick up a test within 10 minutes, the result is Unreachable with the message Agent offline.

SFTP, FTP, file storage, email and API connections are tested from DataLug, even in a Private Agent environment.

A test sent to your agent shows Testing on your agent... on the environment's chip, and its result when the agent answers.

Who is told when a connection fails​

Notifications go out when an environment's first result fails, and when its status changes: from passing to failing, and from failing back to passing. A first result that passes sends nothing. A connection that keeps failing does not send the same notice again on every scheduled run.

When a notification goes out:

  • Your account's Owners and Admins receive an email.
  • The environment's chip on the Connections page shows the new status.
  • The recipient set on the Operational alerts card, while it is on, receives an alert. See Runtime unreachable alerts.
  • Your log destinations receive a connection health event.