Docs › Reference
Troubleshooting
Every error you can hit in the console or a report, and exactly how to fix it.
Console won't connect
| Symptom | Fix |
|---|---|
| Sidebar shows ● Offline / "Not connected to the API" | The API URL is wrong or the API isn't running. Check Settings → API URL. In production it's https://executor-api.planningpowertools.com. |
| Browser console shows a CORS error | The API's CORS_ORIGINS doesn't include the console's origin. It must match exactly (e.g. https://executor.planningpowertools.com). |
| "Not secure" padlock warning | The console subdomain has no SSL certificate. Issue one in your hoster panel (same as the other product subdomains). |
PIN problems
| Symptom | Fix |
|---|---|
| "PIN missing or wrong" when saving/approving | A PIN is set on the server. Enter it when prompted (or under Settings → Supervisor PIN). The PIN is memory-only, so you re-enter it once per session. |
| "Current PIN is wrong" when setting a new PIN | Enter the existing PIN as Current PIN. If the server has POWEREXECUTOR_PIN set, that's the value. CHANGE_ME counts as no PIN. |
| Forgot the PIN entirely | On the server: docker exec powerexecutor-api sqlite3 /data/powerexecutor.db "DELETE FROM config WHERE key='supervisor_pin_hash';", then set a new one under Settings → Security. |
Report says "The AI model call failed"
| Technical detail contains… | Fix |
|---|---|
| "LLM Provider NOT provided" | Provider is blank. Set Configuration → AI → Provider (e.g. deepseek), or include the prefix in the model (deepseek/deepseek-chat). |
| "invalid api key" / "authentication" | Wrong or missing API key. Re-enter it under Configuration → AI. |
| "model not found" | The model name doesn't exist for that provider. Check the exact id (e.g. deepseek-chat). |
Report says "Cannot connect to the product"
| Technical detail contains… | Fix |
|---|---|
"401 Unauthorized" on /api/auth/login | The product password is wrong. Re-enter it under Configuration → Connection. |
| Login works but signals fail | The brand id is wrong or no brand exists. Set a valid Brand id, or create a brand in the product. |
| Connection refused / timeout | The product isn't reachable. Confirm it's running and on the same network as PowerExecutor. |
Nothing happens on schedule
Check Configuration → Operations: Enabled products must list the product, and the current UTC time must be at/after Cycle hour (UTC) (each product runs once per UTC day). You can always trigger a run manually with Run cycle now.
An action failed and it stopped acting
An irreversible failure trips degraded mode. Read the report's Needs-attention card, fix the cause, then Dashboard → Degraded → Clear.
A report always reflects only the cycle that produced it — unknowns reset at the start of each cycle. If you still see an old error, run a fresh cycle.