Will scaling replica nodes for Redis OSS cause some data loss

No, data written during the scaling process should not be lost. When scaling replica nodes for Redis OSS (Cluster Mode Disabled) clusters, the behavior depends on your Redis version:

For Redis OSS 5.0.5 and above: You can scale clusters with auto failover enabled while the cluster continues to stay online and serve incoming requests. This means your application can continue writing data throughout the scaling process without data loss.

For Redis OSS version 4.0.10 and below: You may notice a brief interruption of reads and writes from the primary node while the DNS entry is updated. However, this interruption is very brief (a few seconds), not the entire duration of the scaling process.

During the scaling process, here's what happens:

1. A new cluster with the new node type is created
2. Cache data is copied from the existing cluster to the new cluster
3. Once the copy is complete, reads and writes are served using the new cluster
4. The endpoints remain the same, so no application changes are needed

The key point is that while the data copying may take time (like your 10-minute example), the actual interruption to read and write operations is only a few seconds during the DNS update, not the entire scaling duration. For newer Redis versions (5.0.5+), the cluster remains online throughout, so writes continue to be accepted and no data should be lost.
Sources
Scaling replica nodes for Valkey or Redis OSS (Cluster Mode Disabled) - Amazon ElastiCache
Scaling for Valkey or Redis OSS (Cluster Mode Disabled) clusters - Amazon ElastiCache

Hello.

I have previously conducted tests on ElastiCache Redis regarding write operations during scaling up.
While online scaling is indeed possible, I observed logs showing several write attempts failing, so I believe the number of errors will increase in environments with a high volume of writes.
Since no data was lost, I believe this shouldn't be a major problem if the application is designed to allow for error handling retries.
Also, as mentioned in the following document, it is recommended to perform the operation during times when there is less writing activity, if possible.
https://docs.aws.amazon.com/AmazonElastiCache/latest/dg/redis-cluster-vertical-scaling.html

While the vertical scaling process is designed to remain fully online, it does rely on synchronizing data between the old node and the new node. We recommend that you initiate scale up/down during hours when you expect data traffic to be at its minimum.