During routine maintenance on our database system, an isolated node unexpectedly rejoined the cluster, causing intermittent service interruptions to customers.
We quickly followed established procedures to resolve the issue and successfully restored full functionality.
Later on we noticed that some customers using the API still experienced issues validating their API key. This was also due to data not being replicated properly to all nodes of the database cluster, yet. We addressed the issue by making the API key resolution process more robust in such cases.