Help Centre

Appearance

Troubleshooting ​

Quick Fixes ​

Page Won't Load ​

Symptoms:

  • Blank page or infinite loading spinner
  • Page loads but shows no data
  • "Something went wrong" error

Solutions:

  1. Refresh the page (Ctrl+R / Cmd+R)

    • Many transient issues resolve with a simple refresh
    • Clears cached client-side data

  2. Clear browser cache and cookies

    • Open Developer Tools (F12)
    • Right-click the refresh button → "Empty Cache and Hard Reload"
    • Or: Browser settings → Clear browsing data → Cached images and files

  3. Try incognito/private mode

    • Helps identify if browser extensions are causing issues
    • Bypasses cached authentication state

  4. Check browser console for errors

    • Open Developer Tools (F12) → Console tab
    • Look for red error messages
    • Copy the full error message for support tickets

  5. Verify you're on a supported browser

    • Chrome 90+, Firefox 88+, Safari 14+, Edge 90+
    • Older browsers may not support modern JavaScript features

Authentication Issues ​

Can't Log In ​

Symptoms:

  • Login button doesn't work
  • Redirected to error page after clicking "Sign In"
  • "Invalid session" or "Authentication failed" message

Solutions:

  1. Clear browser cookies

    • Authentication state is stored in cookies
    • Settings → Privacy → Clear cookies for signalbreak.vercel.app

  2. Disable browser extensions

    • Ad blockers and privacy tools can interfere with OAuth flows
    • Try disabling extensions or use incognito mode

  3. Check if your organisation uses SSO

    • SignalBreak supports Single Sign-On for enterprise customers
    • Contact your IT admin if you're unable to access via standard login

  4. Verify your email address is confirmed

    • Check your email for a verification link
    • Resend verification email from login page

Session Expired / Logged Out Unexpectedly ​

Symptoms:

  • Suddenly logged out while working
  • "Session expired" message
  • Redirected to login page

Solutions:

  1. Log back in

    • Sessions expire after 7 days of inactivity for security
    • You'll need to re-authenticate

  2. Check if you're logged in on multiple tabs

    • Logging out in one tab logs you out everywhere
    • This is expected behaviour for security

  3. Disable "Clear cookies on exit"

    • Browser settings → Privacy → Disable "Clear cookies when browser closes"
    • Otherwise you'll need to log in every time you open your browser

Data Loading Issues ​

Workflows Not Appearing ​

Symptoms:

  • Workflows page shows "No workflows registered yet"
  • But you've definitely created workflows
  • Or: Some workflows missing from the list

Solutions:

  1. Check you're viewing the correct filter

    • Look for "Filter by Model" dropdown
    • Ensure it's set to "All workflows" not a specific model

  2. Verify you're in the right tenant

    • If you have access to multiple organisations, check you're viewing the correct one
    • Top-right corner shows current organisation name

  3. Check if workflows were deleted

    • Workflows can be deleted by any team member with access
    • Check with colleagues if you're on a shared account

  4. Refresh the page

    • Sometimes cached data doesn't update immediately
    • Hard refresh (Ctrl+Shift+R / Cmd+Shift+R)

Provider Health Shows "Unknown" ​

Symptoms:

  • Provider health indicator shows ⚪ Unknown
  • "Last checked: never" or very old timestamp

Solutions:

  1. Wait for the next polling cycle

    • Providers are polled every 5 minutes
    • If you just added a provider, wait 5 minutes and refresh

  2. Check if the provider has a status page

    • Some providers (Tier 3/4) don't have official status pages
    • SignalBreak can't monitor providers without public status endpoints

  3. Verify network connectivity

    • If SignalBreak's polling service can't reach the status page, status will show "Unknown"
    • This is usually temporary (network issues, status page maintenance)

Workflow Creation Failures ​

"Cannot remove the last primary provider" ​

Error message:

Cannot remove the last primary provider for this workflow

Cause: Every workflow must have at least one primary provider binding. You're trying to delete or change the role of the only primary provider.

Solution:

Option 1: Add a new primary provider first

  1. Click "Add Fallback" in Provider Bindings section
  2. Select a different provider
  3. Toggle its role to Primary
  4. Now you can delete or change the old primary provider

Option 2: Change a fallback to primary

  1. Find an existing fallback provider binding
  2. Click the role toggle to change it to Primary
  3. Now you can delete or change the old primary provider

"No models enabled for this provider" ​

Error message (in model dropdown):

No models enabled for this provider.
Enable models on the Models page first.

Cause: The provider you selected doesn't have any models enabled in your account configuration.

Solution:

  1. Navigate to Models page

    • Dashboard → Models (in navigation menu)

  2. Find the provider in the list

  3. Enable at least one model:

    • Click "Enable" on the model(s) you want to use
    • Choose models that match your workflow's AI capability

  4. Return to workflow and try again

    • The model dropdown should now show your enabled models

"Provider Configuration Required" Warning ​

Symptoms:

  • Orange warning banner on workflow detail page
  • "This workflow needs at least one AI provider configured"

Cause: You created a workflow but haven't added any provider bindings yet.

Solution:

  1. Click "Add Provider Binding" button in the warning banner

  2. Select a provider from your enabled providers list

    • If no providers appear, visit Dashboard → Providers and enable at least one

  3. Choose a model (if available)

  4. Save the binding

Why this matters: SignalBreak can't monitor workflows without provider bindings—it doesn't know which providers to track for this workflow.

Risk Assessment Issues ​

Risk Score is 0 Despite Having Workflows ​

Symptoms:

  • Dashboard shows Risk Score: 0 / Green status
  • But you have workflows configured

Cause: Risk scores are calculated from scenario impacts. If you haven't executed any scenarios yet, your risk score will be 0.

Solution:

  1. Create scenarios

    • Navigate to Dashboard → Scenarios
    • Create scenarios modelling potential provider disruptions
    • Example: "OpenAI GPT-4 Outage", "Anthropic Rate Limiting"

  2. Execute scenarios

    • Click "Execute" on each scenario
    • SignalBreak analyses which workflows would be affected
    • Creates scenario impacts with severity ratings

  3. Risk score updates automatically

Scenario Execution Shows "No Workflows Affected" ​

Symptoms:

  • Execute a scenario like "OpenAI Outage"
  • Result: "No workflows affected"
  • But you know you have workflows using OpenAI

Cause: Your workflows aren't bound to the provider in the scenario, or the provider bindings don't have models assigned.

Solution:

  1. Check workflow provider bindings

    • Open each workflow detail page
    • Verify Provider Bindings section shows the correct provider
    • Ensure the provider matches the scenario's target provider

  2. Verify models are assigned

    • Each binding must have a model selected (not just a provider)
    • Scenarios match on provider + model, not just provider name

  3. Re-execute the scenario after fixing bindings

Common Error Messages ​

"Tenant or user not found" ​

Cause:

  • Your session is invalid or expired
  • You're trying to access a resource that doesn't belong to your organisation

Solution:

  1. Log out and log back in
  2. Verify you're accessing the correct organisation
  3. If problem persists, contact support—your account may need to be provisioned

"Validation error: At least one field must be provided for update" ​

Cause: You tried to save a workflow without changing any fields (PATCH request with empty body).

Solution:

  • Make at least one change to the workflow before saving
  • If you don't want to make changes, just close the edit form

"Failed to fetch" ​

Cause:

  • Network connectivity issue
  • API endpoint temporarily unavailable
  • CORS error (if accessing from non-standard domain)

Solution:

  1. Check your internet connection
  2. Refresh the page and try again
  3. If using VPN, try disconnecting temporarily
  4. Check browser console for detailed error message

Known Limitations ​

Bulk Import Doesn't Validate Duplicates ​

Status: Active Tech Debt (TECH-DEBT-001)

Description: The bulk CSV import wizard doesn't run fuzzy matching against existing products. You could accidentally import duplicate products with slight spelling variations.

Workaround:

  1. Export your existing products first
  2. Manually check for duplicates before importing
  3. Use the single-add workflow for individual products (has fuzzy matching)

Timeline: Fix planned before general availability launch

Claude Extractor May Miss Some Provider Changes ​

Status: Active Tech Debt (TECH-DEBT-008)

Description: Signal intelligence extraction can fail if Claude refuses to process content or returns malformed JSON.

Impact:

  • Some provider changes detected but not classified
  • Missing signal intelligence for affected providers

Workaround:

  • Check provider changelogs manually for critical providers
  • Cross-reference with provider status page

Timeline: Fix in progress

Performance Issues ​

Page Loads Slowly ​

Symptoms:

  • Pages take >5 seconds to load
  • Dashboard feels sluggish
  • Workflow list takes a long time to appear

Solutions:

  1. Check your network speed

    • Open speedtest.net and verify you have >5 Mbps
    • Slow networks will make any web app feel slow

  2. Reduce browser tabs

    • Too many open tabs consume memory
    • Close unused tabs, especially video streaming sites

  3. Clear browser cache (paradoxically, old cache can slow things down)

    • See "Clear browser cache" steps above

  4. Check for browser extensions causing performance issues

    • Disable extensions one by one to identify culprit
    • Common offenders: grammar checkers, password managers

  5. Try a different browser

    • Chrome generally has best performance for React applications
    • Firefox and Safari are also well-supported

API Requests Timing Out ​

Symptoms:

  • "Request timeout" errors
  • Loading spinners never finish
  • Console shows 504 Gateway Timeout errors

Causes:

  • Database query taking too long (e.g., complex governance report generation)
  • Provider status page taking too long to respond during health polling
  • Serverless function cold start (first request after inactivity)

Solutions:

  1. Wait and retry

    • First request after inactivity can be slower (cold start)
    • Subsequent requests will be faster

  2. Reduce query complexity

    • If generating a report with many workflows/scenarios, try filtering to fewer items
    • Large evidence packs (100+ pages) can take 30-60 seconds to generate

  3. Check if there's a platform incident

    • Visit Dashboard → System Status
    • Or check status.vercel.com for Vercel platform issues

Still Stuck? ​

Before Contacting Support ​

Please gather the following information to help us resolve your issue quickly:

  1. Browser and version

    • Example: "Chrome 120.0.6099.129"
    • Help → About Chrome (or equivalent)

  2. What were you trying to do?

    • Specific steps you took before encountering the issue

  3. Error message (full text)

    • Copy the exact error message if one appeared
    • Screenshot if possible

  4. Browser console output

    • Open Developer Tools (F12) → Console
    • Screenshot any red error messages
    • Copy full error text

  5. When did it start?

    • First occurrence vs. ongoing issue
    • Does it happen every time or intermittently?

Contact Support ​

For critical issues affecting your production workflows:

  • Mark your email subject as [URGENT]
  • Include tenant ID and affected workflow IDs
  • Describe business impact