User Manual MongoDB Atlas Service Discovery Back to User Manual
This guide separates discovery problems from database connection problems. Atlas API Keys and Service Accounts control discovery. Atlas database usernames and passwords control database access.
When one or more credentials fail, the MongoDB Atlas root item shows a recovery action. Healthy resources found through other credentials remain visible.
| Recovery action | What to do |
|---|---|
| Click here to retry | Use this for a temporary network, service, or rate-limit failure. |
| Click here to revisit credentials | Open credential management when a credential is invalid, lacks permission, or requires a more detailed review. |
In credential management, use Retry for one credential or Retry all for every credential. A retry fetches fresh resource information and rechecks Service Account access.
Check the following:
Add another credential when the resources are in a different Atlas organization.
Review the Public Key and Private Key for an API Key, or the Client ID and Client Secret for a Service Account. Then update the credential or add it again.
An Atlas 401 response usually means the credential cannot be authenticated. An Atlas 403 response usually means Atlas accepted the credential but the credential lacks access to the requested resource. Both cases can also require changes to Atlas network access rules.
Use Open in MongoDB Atlas from credential management to review the affected API Key or Service Account. The Atlas API access documentation describes API Key and Service Account setup.
Atlas can restrict requests by source IP address. If Atlas reports that the current address is not allowed, add the appropriate address or network range in Atlas, then use Retry.
The required rule depends on the type of request:
| Request | Access to check |
|---|---|
| Resource discovery through the Atlas Admin API | The network and access restrictions that apply to the API credential. |
| Database connection to a cluster | The Atlas project IP access list and any private networking configuration for the cluster. |
See Configure IP Access List Entries in the Atlas documentation for the current Atlas configuration steps.
Check the cluster status first. A paused, creating, updating, repairing, or deleting cluster is not ready for a connection. Resume the cluster or wait for the Atlas operation to complete.
If the cluster is ready but connection fails, check:
A TLS handshake error indicates that the connection did not complete at the transport layer. It does not, by itself, prove that the database username or password is incorrect. Review the cluster state, network path, and TLS configuration before replacing database credentials.
Atlas can temporarily reject or delay requests because of a rate limit, a network interruption, or a service failure. Wait briefly and use the retry action. Retrying a single credential does not recheck the other configured credentials.
When reporting a problem, include:
Never include Private Keys, Client Secrets, database passwords, access tokens, or full connection strings in a support request.