Astra DB database guardrails and limits

DataStax Astra DB includes guardrails and sets limits to ensure good practices, foster availability, and promote optimal configurations for your databases.

Astra DB offers a $25.00 free credit per month, allowing you to create an Astra DB database for free. Create a database with just a few clicks and start developing within minutes.

Each plan includes a $25.00 free credit per month. The $25 credit is good for approximately 30 million reads, 5 million writes, and 40GB of storage per month on a serverless database.

Limited access to administrative tools

Because Astra DB hides the complexities of database management to help you focus on developing applications, Astra DB is not compatible with DataStax Enterprise (DSE) administrative tools, such as nodetool and dsetool.

Use the DataStax Cloud console to get statistics and view database and health metrics. Astra DB does not support access to the database using the Java Management Extensions (JMX) tools, like JConsole.

Simplified security without compromise

Astra DB provides a secure cloud-based database without dramatically changing the way you currently access your internal database:

New user management flows avoid the need for superusers and global keyspace administration in CQL.
Endpoints are secured using mutual authentication, either with mutual-TLS or secure tokens issued to the client.
TLS provides a secure transport layer you can trust, ensuring that in-flight data is protected.
Data at rest is protected by encrypted volumes.

Additionally, Astra DB incorporates role-based access control (RBAC).

See Security guidelines for more information about how Astra DB implements security.

Replication within regions

Each Astra DB database uses replication across three availability zones within the launched region to promote uptime and ensure data integrity.

Serverless database limits

The following limits are set for serverless databases created using Astra DB. These limits ensure good practices, foster availability, and promote optimal configurations for your database.

Databases

There is a limit of 5 databases per organization. If you need to increase the number of databases for your organization, contact DataStax Support.

Columns

Parameter	Limit	Notes
Size of values in a single column	10 MB	Hard limit.
Number of columns per table	75	Hard limit.

Parameter

Limit

Notes

Size of values in a single column

10 MB

Hard limit.

Number of columns per table

Hard limit.

Tables

Parameter	Limit	Notes
Number of tables per database	200	A warning is issued when the database exceeds 100 tables.
Table properties	Fixed	All table properties are fixed except for Expiring data with time-to-live.

Parameter

Limit

Notes

Number of tables per database

200

A warning is issued when the database exceeds 100 tables.

Table properties

Fixed

All table properties are fixed except for Expiring data with time-to-live.

Secondary indexes and materialized views are not available for serverless databases. Our team is working to offer materialized views for serverless databases soon.

Workloads

Astra workloads are limited to 4096 ops/sec by default. If you see a "Rate limit reached" error in your application and want your limit raised, please open a support ticket.

Storage-Attached Indexing (SAI) limits

The maximum number of SAI indexes on a table is 10. There can be no more than 50 SAI indexes in a single database.

Automated backup and restore

Serverless databases created using Astra DB are automatically backed up once an hour. Snapshots are stored for 20 days, providing a range of options for which you can restore your database, if necessary.

For databases with multiple datacenters in various regions, backup is managed for only the original datacenter selected when creating your database. Restoring from backup removes all other datacenters, restoring to the primary datacenter, and adding the other datacenters with a restored version of the primary datacenter.

If the database was terminated, all data is destroyed and is unrecoverable.

If data is accidentally deleted or corrupted, contact DataStax Support to restore data from one of the available snapshots. This window ensures that the data to restore exists as a saved snapshot.

When restoring data, DataStax Support allows you to restore data on the same cluster if the table is dropped, and data is restored in the same cluster using the specified backup.

Classic database limits

The following limits are set for classic databases created using Astra DB. These limits ensure good practices, foster availability, and promote optimal configurations for your database.

Columns

Parameter	Limit	Notes
Size of values in a single column	5 MB	Hard limit.
Number of columns per table	50	Hard limit.

Parameter

Limit

Notes

Size of values in a single column

5 MB

Hard limit.

Number of columns per table

Hard limit.

Tables

Parameter	Limit	Notes
Number of tables per database	200	A warning is issued when the database exceeds 100 tables.
Table properties	Fixed	All table properties are fixed except for Expiring data with time-to-live.
Secondary index	1	For classic databases, limit is per table.
Materialized view	2	Limit is per table. A warning is issued if the materialized view will create large partitions.

Parameter

Limit

Notes

Number of tables per database

200

A warning is issued when the database exceeds 100 tables.

Table properties

Fixed

All table properties are fixed except for Expiring data with time-to-live.

Secondary index

For classic databases, limit is per table.

Materialized view

Limit is per table. A warning is issued if the materialized view will create large partitions.

Workloads

Astra DB workloads are limited to 4096 ops/s/coordinator by default. If you see a "Rate limit reached" error in your application and want your limit raised, please open a support ticket.

Storage-Attached Indexing (SAI) limits

The maximum number of SAI indexes on a table is 10. There can be no more than 100 SAI indexes in a single database.

Automated backup and restore

Classic databases created using Astra DB are automatically backed up every four hours. The latest six snapshots are stored, providing flexibility in which point in time you can restore to, if necessary.

If the database was terminated, all data is destroyed and is unrecoverable.

If data is accidentally deleted or corrupted, contact DataStax Support within 12 hours to restore data from one of the available snapshots. This window ensures that the data to restore exists as a saved snapshot.

When restoring data, DataStax Support allows you to restore data on the same cluster if the table is dropped, and data is restored in the same cluster using the specified backup.

Cassandra Query Language (CQL)

At this time, user-defined functions (UDFs) and user-defined aggregate functions (UDAs) are not enabled.

Parameter Limit Notes

Parameter	Limit	Notes
Consistency level	Fixed	Supported consistency levels: Reads: Any supported consistency level is permitted Writes: LOCAL_QUORUM and LOCAL_SERIAL
Compaction strategy	Fixed	UnifiedCompactionStrategy is a more efficient compaction strategy that combines ideas from STCS (SizeTieredCompactionStrategy), LCS (LeveledCompactionStrategy), and TWCS (TimeWindowCompactionStrategy) along with token range sharding. This all-inclusive compaction strategy works well for all use cases.
Lists	Fixed	Cannot `UPDATE`, `INSERT`, or `DELETE` a list value by index because Astra DB does not allow list operations which perform a read-before-write.
Page size	Fixed	The proper page size is configured automatically.
Large partition	Warning	A warning is issued if reading or compacting a partition that exceeds 100 MB.

Consistency level

Fixed

Supported consistency levels: Reads: Any supported consistency level is permitted Writes: LOCAL_QUORUM and LOCAL_SERIAL

Compaction strategy

Fixed

UnifiedCompactionStrategy is a more efficient compaction strategy that combines ideas from STCS (SizeTieredCompactionStrategy), LCS (LeveledCompactionStrategy), and TWCS (TimeWindowCompactionStrategy) along with token range sharding. This all-inclusive compaction strategy works well for all use cases.

Lists

Fixed

Cannot UPDATE, INSERT, or DELETE a list value by index because Astra DB does not allow list operations which perform a read-before-write.

Page size

Fixed

The proper page size is configured automatically.

Large partition

Warning

A warning is issued if reading or compacting a partition that exceeds 100 MB.

cassandra.yaml

If you are an experienced Cassandra or DataStax Enterprise user, you are likely familiar with editing the cassandra.yaml file. For Astra DB, the cassandra.yaml file cannot be configured.

The following limits are included in Astra DB:

// for read requests
        page_size_failure_threshold_in_kb =  512
        in_select_cartesian_product_failure_threshold =  25
        partition_keys_in_select_failure_threshold = 20
        tombstone_warn_threshold = 1000
        tombstone_failure_threshold = 100000

// for write requests
        batch_size_warn_threshold_in_kb = 5
        batch_size_fail_threshold_in_kb = 50
        unlogged_batch_across_partitions_warn_threshold = 10
        user_timestamps_enabled = true
        column_value_size_failure_threshold_in_kb = 5 * 1024L
        read_before_write_list_operations_enabled = false

// for schema
        fields_per_udt_failure_threshold = 30
        collection_size_warn_threshold_in_kb =  5 * 1024L
        items_per_collection_warn_threshold =  20
        columns_per_table_failure_threshold = 50 (Classic) or 75 (Serverless)
        secondary_index_per_table_failure_threshold = 1
        tables_warn_threshold = 100
        tables_failure_threshold = 200

// for node status
        disk_usage_percentage_warn_threshold =  70
        disk_usage_percentage_failure_threshold =  80
        partition_size_warn_threshold_in_mb =  100

// SAI Table Failure threshold
        sai_indexes_per_table_failure_threshold = 10