SharkMQ Troubleshooting
| Symptom | Likely cause and fix |
|---|---|
InvalidClientTokenId or SignatureDoesNotMatch | Wrong access key id or secret, or a secret that was rotated. Check the key in Users & keys; rotate it if the secret is lost. |
QueueDoesNotExist | The queue was never created, or the URL names another broker. Use GetQueueUrl. |
UnsupportedOperation mentioning the JSON protocol | Your AWS SDK is older than 2023. Update it. |
AMQP ACCESS_REFUSED | Wrong user or secret, or the virtual host is not your broker id. |
AMQP NOT_ALLOWED / access to vhost refused | The virtual host in your URL is wrong. |
| Connections refused | The connection limit is reached, or the broker is suspended (check the portal). |
| Messages arrive twice | Expected after a failure (at-least-once delivery). Make processing safe to repeat, for example by remembering message ids. |
| Messages disappear without being read | The queue reached its size limit (oldest dropped), or the retention time passed. |
| Sending is slow | Each confirmed send waits about 0.15 s for a second building. Send in batches, or keep several sends in flight. |
| A message keeps coming back | Your consumer does not delete (SQS) or acknowledge (AMQP) it. Add a dead-letter queue so it stops after N tries. |
Still stuck? Open a support ticket with your broker id and the error message.