What it is: Atlas now automatically alerts you and recovers from failures when exporting logs to an external destination (like S3, Datadog, Splunk, or an OTel collector). If log delivery fails, you'll get an alert and an Activity Feed event, and once the connection is restored, queued logs are automatically replayed to your destination.

Who it's for: Atlas customers using External Log Sinks to export mongod, mongos, or audit logs to third-party observability or storage tools.

Why it matters: Previously, if a log sink integration failed (due to a misconfiguration, expired credentials, or a downed destination), logs would queue silently with no automatic recovery. Now you're notified as soon as something goes wrong, and once the issue is resolved, queued logs are delivered automatically, so you don't lose visibility or need to manually trigger replay.

How to get started: These alerts are enabled by default for all projects, no setup required. You can view failure and recovery events in the Activity Feed, or customize notification behavior in the Atlas Alerts UI.