Common issues

The problems users hit most often, and the first thing to try for each.

"I can't log in"

Right URL? Confirm you're hitting your actual Flero instance, not a stale bookmark.
Right email? A typo is the most common cause.
Password manager filling wrong credentials? Try in private/incognito mode.
MFA loop? If you keep being asked for an MFA code that doesn't work, your authenticator app's time may have drifted. Resync the app's clock.
Approval pending? If your account was just created, an admin may need to approve. Check your email or ask them.
SSO redirect? If your org has SSO, the login form may redirect you to your IdP, make sure you can sign in there.

If none of these: ask an admin to check your account in the user list. They can see the last-login timestamp and any auth errors.

"My workflow won't save"

The Save button is greyed out or shows an error banner.

Validation errors. Look for nodes with a red border on the canvas. Hover for the error chip. Common: missing required fields, invalid expressions.
No active workspace / read-only mode. Look at the top sidebar, are you in a workspace? Is the workspace archived?
Concurrent edit conflict. Someone else just saved. Refresh and retry.
Network blip. Check the toast notification; click Retry if present.

"I clicked Run but nothing happens"

Check:

Workflow saved? Save first.
Validation passed? No red nodes.
Workspace at quota? A workspace that's hit its monthly execution cap can't start runs. Check Settings → Billing & usage.
Worker pool down? For self-hosted: check the admin dashboard for worker health.

If the run started but is stuck on pending, the queue is backed up, wait, or scale workers.

"A node is failing in production but works in sandbox"

The classic. Most likely causes:

Sandbox mocks differ from real responses. Pin the actual production response as your sandbox mock so the two converge.
Real credential is expired or wrong scope. Test the credential in Settings → Credentials.
Rate limit in real, mock returns instantly in sandbox. Add retries with backoff.
Data shape varies in production. A field that's always present in your mocks is sometimes missing in real data. Add a Schema Validator earlier in the chain.

See Sandbox mode, troubleshooting.

"I get `[EXPR-ERROR: ...]` in my output"

Read the marker, it's deliberately informative. Most common:

Typo in $node["..."]. Copy the exact name from the canvas.
Forgot the $node["..."] wrapper around a bare reference like start.output.x.
Referenced a downstream node, or a node that didn't run (wrong If branch).

Full guide: Debugging expressions.

"Webhook URL returns 410 Gone"

Workflow status is draft or inactive. Set to active from the workflow editor toolbar.

"I can't see my workflow / credential / template"

Wrong workspace? Check the sidebar workspace name.
Filter / search active? Clear filters; check the empty-state copy.
Archived? Open the Archive tab (workflows) or filter scope (credentials).

"Sandbox toggle isn't where the docs say it is"

The toggle is in the bottom execution chrome (not the toolbar). If the chrome is hidden (very narrow viewport), expand it via the chevron at the bottom of the screen.

"My approval is stuck, nobody's responding"

Approver doesn't know. Check their notification settings; manually nudge them.
Approver doesn't have access. Look at the approval card, is the named approver still a workspace member?
Public approval link emailed but bounced. Check the email send log.
Token expired. If it's been weeks, regenerate by canceling and re-running.

Cancel the approval from /approvals to unblock the workflow.

"Costs are higher than expected"

AI tokens. Open Settings → Billing & usage → AI tokens. Use the Token Cost Guard node on hot workflows.
Excessive retries. Check workflows with Max retries set high on paid APIs.
Polling triggers. Every poll is a service API call. Increase the poll interval for non-time-sensitive triggers.
Sandbox runs hitting real services. Verify no nodes have Disable sandbox for this node toggled on by mistake.

"I made a change and broke production"

Don't panic.
Workflow toolbar → Version history → Restore previous version. Each save is a checkpoint.
Once restored, investigate the difference at leisure.

"My data isn't showing up in `/logs` or `/monitoring`"

Retention window. Older runs are purged. Check Settings → Data & privacy → Retention.
Sandbox flag. Sandbox runs are hidden from monitoring by default; toggle Include sandbox.
Permission scope. Viewers see only what they have access to.

When to escalate

If you've tried the obvious fixes and the issue persists, file a bug with:

Execution ID (if applicable).
Steps to reproduce.
Expected vs actual behaviour.
Whether the workspace is shared (so support can investigate without screen-share if needed).
Screenshot / log of the error.

Found something out of date? This page lives in the Flero docs content set.