We use Astra DB to enable our Graph Database functionality. This graph database is the core of our e-commerce delivery planning and execution systems. This system enables the business to provide accurate ETAs to customers. It also enables the business to grow our national delivery network simply via configuration. Routes and their capabilities are individually configured to account for the network capacity and courier capabilities around different geographic areas.
Pros
Migration
API Integration
Visibility
Support
Cons
Support - Turnaround time was slow.
Likelihood to Recommend
Anyone looking for a hosted solution of Cassandra DB will find a good offering with AstraDB. It provides the scalability of Cassandra with added security, permissions and visibility. As a user you forget there is a cluster behind the scenes.
Astra DB is cloud-native, deployment and management are simplified.
Astra DB comes equipped with robust developer tools, including CQL (Cassandra Query Language), REST APIs, and GraphQL support.
Astra DB's Storage Attached Index(SAI) allows for efficient indexing directly on disk, making it much easier to perform complex queries across large datasets
Likelihood to Recommend
For vector search capabilities where you need some powerful querying capability like CQL, Astra DB is the solution. Also Astra DB suits well where someone wants to build a RAG setup. Its cloud-native design and distributed architecture make Astra DB a great fit for companies operating across multiple regions or requiring high availability.
We mainly use Astra DB Vector databases for our internal and customer support chatbots. The internal chatbot uses a database with more critical data and is therefore, separated from the customer facing chatbot.
Pros
Very fast vector search
Easily configurable
Great and very responsive customer support
Cons
Copying of databases and relocating them is not possible
Likelihood to Recommend
Especially the personal customer support over Slack is very helpful and this why I would always recommend Astra DB to anyone starting with RAG and LLMs.
We use Astra DB to power core components of our application including our Feed, Chat & AI Vector search. All our user feeds and chat data and stored and loaded from Astra DB in real time.
The vector search is used to power our recommendations engine.
Pros
It's very resilient and scalable, no downtime and no issues scaling up to meet our needs.
Low latency reads and writes
Cost effective - The on demand model worked out cheaper than running our own clusters
Great support for any of our questions or issues
Cons
It can be a little difficult to tell how many credits are used by each database by just seeing reads & writes.
Nothing else particularly, we haven't run into any real problems.
Likelihood to Recommend
The perfect use case for Astra DB is if you have an existing Cassandra setup and would like to move to a cost-effective managed service. Even if you aren't using Cassandra yet, if you have workloads with very high writes that need low latency, it's a great choice. Any no-sql key-value or wide column workloads.
It's also great for vector search.
It's not great for very relational data models with a lot of joins, etc.
We connected Astra DB as a vector database to our backend. The main use case for it is that we store the embeddings of our main data entities. In Restworld, we have something similar to a marketplace made of job opportunities and job seekers. We use text embedding models to create vectors out of them and we exploit Astra DB vector database capabilities to power our recommendation and personalization algorithms.
Pros
Scalability
Flexibility
Enterprise support
Cons
Simplification of API choices
Likelihood to Recommend
Well suited: - Retrieval augmented generation. It is best combined with LLM solutions for generative AI applications - Recommendation algorithms. Whenever you need to show personalized content to your users, using vector search is an easy and effective solution to implement - Semantic search. Similar to previous, just with a different vie angle to your data.
Astra DB handles all of the databases queries our product generates. It is the backbone of our product. Every data operation is done through Astra DB. All in all relief for developers.
Pros
Data management becomes easy.
Handles volumes well.
Migration of the data was not as difficult as it was supposed to be.
Cons
API features like LWT need optimization
Documentation could also be focused towards novice user.
Pricing should have been less or at par with the competition in the market.
Likelihood to Recommend
While handling data at a large scale is where Astra DB really shines. The scalability and maintenance department takes 5 stars whereas the security and the ease of learning department falls short. But the overall performance is truly on the positive side of the scale.
We run an advertising company with thousands of websites. A lot of data is generated from each of the auctions for each ad unit on all of those websites, and we need fast response times to make near-instant bidding decisions for those auctions. AstraDB helps store all of that data in a fast, efficient, and easy way to use.
Pros
Great documentation
Easy to use. If you've used a cassandra database before, Astra will be second nature
Great customer support
For languages not officially supported, there is a sidecar application which will expose Astra as if it were a native cassandra cluster
Cons
While there are libraries for many popular languages, it is a little rough hooking up an application written in a language not in the list. The ability is there, but docs for the sidecar are hard to find.
Likelihood to Recommend
Astra is great for when you need a quick and affordable cassandra cluster with zero maintenance overhead. Setup is easy, and you can get running in minutes. Response times across the globe are quick. Monitoring your cluster is easy using provided prometheus and grafana graphs. Astra does limit the amount of control you have over your cluster and sharding. (not a ton, but more restrictive than running your own cluster, as expected.) So if you really need fine-tuned control, Astra may not be the best fit. That being said, the point of having a managed cluster is that you do not need to perform all of that management yourself.
Astra DB used for region replication, efficient scaling scale out using serverless architecture and storing our business data efficiently. We also explored their REST API and Document API
Pros
scales to fit your needs
cost saving on infrastructure
prevent application downtime
Cons
realtime data ingestion from different data sources
Likelihood to Recommend
If someone want self managed database service which can scale in scale out effectively and you will be only charged for your usage then astra DB is best suitable
We use Astra DB for storing big chunks of data. Previously, we were storing this data as files and need to keep local copies when the data was modified or updated. By using this product, we were able to reduce the amount of usage of our cloud storage system. It also gives us the capability to keep track of the most recent generated data and allows us to prevent files from being accidentally overwritten, thus losing previously generated information.
Pros
Easy to use with python
Good online support
Good documentation
Cons
Installing the python driver faster
Better interface for connecting the different tables from the DB form the website
Easy way to monitor the status and health of the DB
Likelihood to Recommend
For implementing timeseries generated data with python scripts, for example, data generated by sensors in case the data is generated in a low-frequency example making a post every 0.5 seconds or so. When the data is generated in a higher frequency sometimes the connection can fail, and the data might not be stored correctly into the DB.
