What do you like best about the product?

For our log data aggregation and analysis platform, ScyllaDB offers unparalleled performance and scalability. Real-time processing of millions of log events per second has relied heavily on its capacity to manage high-throughput writes with incredibly low latency. We also value the low operational overhead—compared to other NoSQL databases, ScyllaDB uses less infrastructure because of its auto-tuning features and effective hardware utilization. In particular, the native support for time-based TTLs is useful for automatically controlling log retention. It just operates on a larger scale.

What do you dislike about the product?

ScyllaDB has a learning curve for teams used to Apache Cassandra or more traditional relational databases even if it boasts better performance. For some complex tuning parameters and architectural ideas, such log analytics, optimal performance for particular workloads depends on a better knowledge. Moreover, even if the documentation is usually good, some edge cases—especially those related to large-scale cluster management and repair activities—would benefit from more practical examples. The community and support network, however, are flexible which helps to close the distance.

What problems is the product solving and how is that benefiting you?

ScyllaDB addresses our fundamental need for real-time ingestion, storage, and querying of enormous amounts of log data. With its high throughput and low-latency performance, ScyllaDB easily addresses the write-heavy nature and time-series structure of logs, which traditional databases battled. Its support of wide rows and TTLs helps to effectively retain logs and query without requiring intricate workarounds. All at scale and with a smaller operational footprint, this has allowed us to provide our clients real-time insights, fast search capabilities, and consistent alerting.