The cloud-native cluster system Vitesse is now available in version 20. Along with the update of the project hosted by the Cloud Native Computing Foundation (CNCF), the development team is also delivering the Vitesse Kubernetes Operator to version 2.13.0. Apart from some updates to existing functions and various changes involving breaking changes, most of the innovations are aimed at contributing to a better user experience. Among other things, the release provides extended support for Data Manipulation Language (DML), which opens up additional options for queries.
Advertisement
More control and efficiency with shared data
According to the announcement in the CNCF blog, Vitesse 20 users can count on high adaptability when it comes to queries. This ranges from Vindex notices for shard updates with limits and multi-table updates to advanced delete operations designed to provide more control and efficiency when managing shard data. For example, Vindex hints allow you to have a more targeted impact on shard routing.
SELECT * FROM user USE VINDEX (hash_user_id, secondary_vindex) WHERE user_id = 123;
SELECT * FROM order IGNORE VINDEX (range_order_id) WHERE order_date = '2021-01-01';One of the important changes that users need to pay attention to when moving to the new release is related to the renaming of various metrics that are relevant to the use of the cluster monitoring and recovery component VTORC. For the time being, the old names will remain the same /debug/vars Received, but will be removed in later releases. The following table compares the old and new metric names – including the common Prometheus names they resolve to:
|
Old metric name |
New metric label |
Names in Prometheus |
|
analyze.transform.write |
Write Change Analysis |
vtorc_analyse_change_write |
|
audit.write |
Auditorite |
vtorc_audit_write |
|
search.try |
Attempted searches |
vtorc_discoveries_attempt |
|
search.fail |
hunt failed |
vtorc_discoveries_fail |
|
searches.instance_poll_seconds_more |
DiscoveryInstancePollSeconds Exceeded |
vtorc_discoveries_instance_poll_seasons_exceeded |
|
searches.queue_length |
Search queue length |
vtorc_discoveries_queue_length |
|
searches.recent_count |
Discoveries RecentCount |
vtorc_discoveries_recent_count |
|
Example.Read |
Read examples |
vtorc_instance_read |
|
example.read_topology |
InstanceReadTopology |
vtorc_instance_read_topology |
|
emergency_parent_count |
emergency guardianship calculation |
vtorc_emergency_reparent_counts |
|
planned_parenthood_count |
Planned Parenthood Count |
vtorc_planned_reparent_counts |
|
parent_shard_operation_timing |
RetainShardOperationTimings |
vtorc_reparent_shard_operation_timing_bucket |
New experimental features
For multi-tenant architectures often found in web applications, where each tenant uses its own database with the same schema, the new experimental VReplication feature aims to simplify multi-tenant imports. To ease challenges such as provisioning and scaling in such architectures, a sharded Vitess keyspace is recommended, which serves all clients with a single logical database. The required data imports into the keyspace can be done in Vitess 20 via VReplication MoveTables-Workflow with flags --shards And --tenant implement.
The new functions in Vitesse Kubernetes Operator 2.13.0 are still considered experimental. The operator now offers the option to set up automatic and time-controlled backups. Details about how to handle these security functions are provided Updated User Guide,
More information about the new major release can be found CNCF blog post about the availability of the Vitesse 20. A detailed overview of all innovations Provide release notes on GitHub,
(Map)
