Check why a connector is failing

What you'll achieve

The actual error message behind a connector's FAILED badge — and, where possible, the connector running again.

Before you start

• Reading the error needs no special rights; restarting needs Developer or higher and an environment that allows changes — in a read-only environment only admins can restart.

Steps

1. Click Connectors in the sidebar. A failing connector shows a red FAILED badge with a small red info icon next to it (hover text: "View error trace").
2. Click the info icon. An Error window opens with the full error text — read the first lines; they usually name the cause (an expired credential, a missing topic, an unreachable external system, …).
3. Fix what the error names, if it's yours to fix — e.g. wrong credentials mean updating the connector's configuration via Edit on its page.
4. Click Restart (on the connector's page, or the restart icon in the list's Actions column — hover text "Restart connector").

What success looks like

"Connector restart action completed successfully" appears, and after a few moments the badge returns to RUNNING. If it flips back to FAILED, the underlying cause isn't fixed yet — read the trace again.

Something went wrong?

• No info icon next to FAILED — Confluent reported no error text; restart once, and if it fails again contact the platform team.
• The error names something outside your reach (an external system, permissions you don't control) — pass the error text to the team that owns that system, or to the platform team.
• Restart is greyed out — it's only available for RUNNING or FAILED connectors, and needs edit rights in this environment.

Background — connector statuses and best practices? Read the Connectors guide.