What I like most about ScyllaDB is its performance and efficiency. It handles very high throughput with low latency, which is critical for real time and large scale data workloads. I also appreciate that it stays compatible with the Cassandra ecosystem, so teams can adopt it without major changes. The way it uses hardware resources efficiently makes operations much simpler. Review collected by and hosted on G2.com.

One thing I find challenging with ScyllaDB is the learning curve when it comes to tuning and operating it at scale. While it delivers strong performance, teams still need a good understanding of distributed systems to get the most out of it. Review collected by and hosted on G2.com.

I use ScyllaDB for building high performance backend systems that need large volumes of data with very low latency. I like its performance and latency, as it can process large numbers of read and write operations while keeping the response time fast. I also like its ability to scale easily. Additionally, it's scalable and compatible with Apache Cassandra APIs, which makes integration with tools and systems much easier. Review collected by and hosted on G2.com.

One thing that could be improved in ScyllaDB is the complexity of the initial setup. For beginners, cluster configuration and tuning can take some time and requires prior knowledge. More beginner-friendly, simplified tools and setup, clearer responses for common deployment scenarios, and automated config options could help developers start more easily. I think more friendly guides could make the process much easier for new users. Review collected by and hosted on G2.com.

I like ScyllaDB's impressive performance and efficiency, delivering very low latency even under heavy workloads. Its excellent use of available hardware allows us to handle large datasets with fewer nodes, which helps in reducing operational costs. The seamless compatibility with Apache Cassandra is also a big plus, as it makes adoption easier while we still benefit from significant performance improvements and scalability. I value how it allows us to keep our APIs responsive and services stable during traffic spikes. Review collected by and hosted on G2.com.

One area that could be improved in ScyllaDB is the operational learning curve. While it performs extremely well once properly configured, tuning and managing clusters can require solid distributed systems knowledge. Some ecosystem tools and documentation are also less mature compared to Apache Cassandra, so troubleshooting or advanced setups may take additional effort. Review collected by and hosted on G2.com.

ScyllaDB delivers extremely low latency and high throughput at scale, which helps keep performance predictable even under heavy load. Its shard-per-core architecture and efficient I/O scheduling allow it to fully utilize modern hardware without constant tuning, and that can reduce the number of nodes we need. I also appreciate that it’s largely Cassandra-compatible while being much faster, so you get a familiar data model and ecosystem, along with better tail latencies. Review collected by and hosted on G2.com.

ScyllaDB still comes with a learning curve, particularly when it comes to data modeling and capacity planning. If you get the model or the underlying infrastructure wrong, performance can degrade quickly. It also assumes you have solid hardware and that you make careful choices around storage and memory, so it’s less forgiving than simpler databases. Because of that, it can feel like overkill for smaller or more generic workloads. Review collected by and hosted on G2.com.

ScyllaDB delivers extremely low latency and high throughput at scale, so it keeps performance predictable even under heavy load. Its shard-per-core architecture and efficient I/O scheduling let it fully utilize modern hardware without constant tuning, which reduces the number of nodes we need. I also like that it is largely Cassandra-compatible while being much faster, so you get a familiar data model and ecosystem with better tail latencies. Review collected by and hosted on G2.com.

ScyllaDB still has a learning curve, especially around data modeling and capacity planning; if you get the model or infrastructure wrong, performance can degrade quickly. It also expects solid hardware and careful storage and memory choices, so it is less forgiving than simpler databases and can feel like overkill for small or generic workloads. Review collected by and hosted on G2.com.

I use ScyllaDB for handling large-scale data in high-performance applications. I like its high performance and low latency when handling large amounts of data. It is designed to efficiently use modern hardware, making data processing very fast. Its compatibility with Apache Cassandra is great since it makes it easier for developers to migrate existing systems. I find the shard-per-core architecture, which efficiently utilizes modern multi-core CPUs, to improve performance and resource usage. Its compatibility with Apache Cassandra supports the same CQL (Cassandra Query Language), making migration easier without needing major code changes. The ability to deliver very low latency and high throughput is ideal for real-time applications that process large volumes of data. Review collected by and hosted on G2.com.

One area that could be improved in ScyllaDB is its learning curve for beginners, especially when setting up and managing clusters. The configuration and performance tuning can be complex for new users. Additionally, better documentation and simpler setup tools would make it easier for developers to get started quickly. Review collected by and hosted on G2.com.

I appreciate ScyllaDB for its predictable low latency at scale, with P99 latency staying in the single-digit millisecond range even under heavy load. I like that we can use fewer nodes while achieving the same or better throughput, which makes our cluster much smaller. The ability to store both our key-value and vector workloads in one database is a big plus, eliminating the need for a separate vector database for RAG and semantic caching. The CQL compatibility is beneficial as it uses the same query language and driver model as Cassandra, easing both migration and hiring. The Shard-per-core-design is a standout feature—it scales performance predictably with core additions, reducing surprises under load and minimizing the need for constant tuning to avoid hotspots or mysterious latency spikes. Review collected by and hosted on G2.com.

Vector search is only on ScyllaDB Cloud right now, and we'd like to see it in open-source/self-managed for on-prem and air-gapped use. Data modeling is strict: partition key design and row/partition size really matter, and it's easy to paint yourself into a corner if you don't plan upfront. There's also a learning curve around compaction strategies, consistency levels, and limits (e.g., tables per keyspaces, partition size) that we had to absorb before we felt confident. Review collected by and hosted on G2.com.

I use ScyllaDB for high throughput backend services that require low latency reads and writes at scale. I like its sharding and shard per core architecture. I also appreciate its compatibility with Cassandra, supporting Cassandra drivers and CQL, which my organization was already experienced with. Switching to ScyllaDB was easy because the same queries and driver configurations worked without code changes. It integrates well with Apache Cassandra drivers, monitoring tools like Prometheus, and visualization platforms like Grafana. The initial setup was fairly straightforward, thanks to the available documentation and compatibility with Apache Cassandra. ScyllaDB delivers excellent performance and scalability for high-throughput workloads. Review collected by and hosted on G2.com.

I think the ecosystem and community resources are still smaller, which can be a bit limiting. The configuration also feels a bit manual, especially when it comes to node removal. I feel expanding documentation with more real-world architecture examples, migration guides from Apache Cassandra, and best-practice patterns for scaling and tuning clusters would help developers adopt it faster. A larger community presence with more open discussions, tutorials, and troubleshooting resources would make it easier for teams to resolve issues and learn from practical production use cases. Review collected by and hosted on G2.com.

ScyllaDB stands out for its performance and efficiency when working with large-scale, data-intensive workloads. The architecture is designed to fully utilize modern hardware, which helps deliver low latency and high throughput even under heavy traffic. I also appreciate that it is compatible with Cassandra, which makes migration or integration easier for teams already familiar with that ecosystem. The documentation and technical talks from the community make it easier to understand distributed database concepts and real-world scaling challenges. Review collected by and hosted on G2.com.

Because ScyllaDB is designed for high-scale distributed systems, it can feel complex for beginners who are new to distributed databases. Some concepts around cluster management, tuning performance, and infrastructure requirements may require a learning curve. Smaller teams or projects that do not require extreme scale might find it more advanced than what they need compared to simpler managed databases. Review collected by and hosted on G2.com.

I find ScyllaDB to be really effective for managing the world's largest data migrations with 99.9% data quality and minimal downtime. I really like the High Velocity Migration and scalable metadata storage features. ScyllaDB acts as a staging layer for high volume migrations and offers high speed. It also handles scalable metadata storage efficiently by managing large catalogs for metadata and business rules. It works well for massive throughput and predictive latency. I also appreciate that setting it up on the cloud is easiest, especially using Docker, though it requires knowledge of persistence storage and networking. Review collected by and hosted on G2.com.

I think ScyllaDB's learning curve could be improved. Also, the monitoring can get quite complex. Setting it up on Docker is a bit tricky because you need to know about persistent storage and networking. Review collected by and hosted on G2.com.