Troubleshooting

Find your symptom below. Most problems come down to a permission, a missing webhook or a missing source. If you see a TX code, look it up in Error codes.

Connections

A connection fails to save

TraceX checks credentials with a real read before saving, so a failure means the read didn’t work.

• Database: check the host, port and database name, that the user can log in and SELECT, and that the database accepts connections from the internet (or contact support about network access).
• Log store or Sentry: check the key’s scopes and region on the integration’s page.
• Expired keys: tokens with an expiry date stop working silently. Create a new one.
• The message ends with a TX code (Error codes).

A connected database shows missing tables

The user can only see what it was granted. Grant SELECT on the missing tables or schemas, then reconnect so TraceX re-reads the schema.

Investigations

An investigation found no root cause

The report says what was missing. Usually one of these helps:

• Add the systems involved: the database, the repository, and logs or errors.
• Be specific: an order number, a customer email, a request id, the time it started.
• Add context: runbooks and architecture notes in the Knowledge base, notes about tables in the company profile.
• Correct TraceX on the report: it learns for next time.

The root cause is wrong

Correct it at the bottom of the report. The correction becomes a fact TraceX uses in later investigations. If it misread a table or field, add a note to the company profile so it applies everywhere.

The “Open pull request” button is missing

Fix pull requests need the Pro or Team plan and a GitHub repository. If your organization requires admin approval, ask for a pull request and an admin opens it. See Fixes and pull requests.

Intake

Slack doesn’t answer

• Invite the app to the channel: /invite @TraceX.
• Make sure the message mentions @TraceX.
• Check Integrations → Slack shows the workspace as connected to this project.

Jira, Linear, Zendesk or Intercom issues aren’t picked up

• Check the webhook from the integration’s page is added in the tool, with the events listed on that integration’s docs page.
• Jira: the issue must be assigned to the TraceX account. Without the webhook, TraceX checks for new assignments every minute.
• Linear, Zendesk, Intercom: the label or tag must match exactly.
• Automatic intake is part of the Team plan.

PagerDuty or Opsgenie incidents aren’t investigated

Check the webhook is scoped to the right services, and that the incident’s urgency or the alert’s priority is within what you chose to investigate.

Account and access

“You don’t have a seat” (TX-3003)

The organization has more members than seats. An admin chooses who keeps a seat, or adds seats, on Billing (Seats).

“Your plan’s limit is reached” (TX-3002)

The organization used this month’s allowance. Admins see usage under Settings → Plan and limits and can upgrade on Billing (Plans).

I lost my two-factor device

Sign in with a recovery code. See Two-factor sign-in.

The CLI says the session ended

Run tracex login again. Sessions last 30 days and can be revoked under Profile → Sessions.

Still stuck

Email [email protected] with:

• the TX code and the reference shown with the error, if any,
• the investigation id (like TX48211794554145), if it’s about an investigation,
• what you expected and what happened.