Skip to content

MatrixOne Intelligence 2023 Release Report

December 23, 2023

MatrixOne Kernel Version Update

Upgraded from v23.1.0.1 tov23.1.0.2. For details, please refer to the MatrixOne v23.1.0.2 Release Notes.

Improvements

  • Optimized thequery history implementation mechanism, significantly improving its response speed.
  • Enhanced the display oferror messages when SQL execution fails; the interface will now show more specific error details to help users analyze and troubleshoot.
  • Improved the display ofmonitoring tick marks: the y-axis tick marks are now fixed at 6, with values determined by their relation to higher-level intervals, providing a more elegant display for data monitoring.
  • Optimized theoverall resource allocation of the cloud platform, leading to improved platform performance.

Bug Fixes

  • Fixed an issue withinstance version query errors.
  • Fixed incorrectalerting.
  • Fixed slow responses when viewingsingle SQL details.
  • Fixed an issue where the query editor did not time out to the login page after prolonged inactivity.
  • Fixed unsuccessfuldatabase switching in the query editor in some scenarios.
  • Fixed an issue where executing SQL in the query editor remained in a "running" state indefinitely in some cases.
  • Fixed SQL execution failures in the query editor in certain scenarios.

December 16, 2023

MatrixOne Kernel Version Update

Upgraded from v23.1.0.0 tov23.1.0.1. For details, please refer to the MatrixOne v23.1.0.1 Release Notes.

Improvements

  • Optimized theMatrixOne Intelligence public beta registration process by merging the trial application and registration pages. Users now only need to fill in information once, though approval from the MatrixOne Intelligence team is still required during the beta period. Please be patient.
  • Optimized thecalculation logic for table storage size. We've refined the statistical logic for table storage size to exclude intermediate data from the object storage process during data persistence, including only the compressed size of the final data and its metadata.
  • Optimized thedisplay units for CU (Compute Units). Based on reading habits in Chinese and English, and the display precision of CU for instances and SQL, we've optimized the display for different CU counts to improve readability for users.
  • Improved theresponse accuracy of the AI intelligent assistant.

Bug Fixes

  • Fixed an issue whereinstance storage was 0 in some scenarios.
  • Fixed an issue where filtering for "failed" status SQL in the query history function returned no results.
  • Fixed an incorrect page redirection when theactivation email expired after user registration.
  • Fixed incorrecttransaction monitoring values.
  • Fixed an issue where SQL users with normal permissions could not use SQL workbooks.
  • Fixed an issue whereinstance storage monitoring values were inaccurate in some cases.
  • Fixed an issue whereAI intelligent assistant reply document links were invalid.
  • Fixed incorrect pagination after changing the list aggregation granularity in the billing function.

November 13, 2023

Features

[Feature 1] Supports Billing Functionality

With version v0.10, we officially introduced robustbilling functionality. This means paid instances are now available to all users. You can top up your platform account viaoffline recharge or by usingvouchers we provide. As long as your account has funds, you can create and use Serverless paid instances—a powerful instance with HTAP capabilities, automatic scaling of compute resources, and isolated read/write resources, designed to enhance your experience.

Specifically, the billing features includeAccount Overview, Transaction Details, Consumption Trends, and Billing Details, allowing you to clearly understand how every penny is earned and spent, and how each cost is generated. For example:

  • How much was spent each month over the past year, and how much was paid by cash versus vouchers.
  • Which were the top 3 instances by consumption over the past 3 months.
  • The amount of each cash top-up or deduction, and the balance at that time.
  • The effective period, face value, and usage records of each voucher held.
  • The hourly cost of a specific instance and its billing logic.
    [Feature 2] MO Robot Assistant Launched!

Leveraging MO's built-in vectorization capabilities, theMO Robot Assistant has officially launched! This is a great realization of MO's vectorization + AI application. You can wake it up by clicking on it in the bottom right corner of the cloud platform. Currently, it's still a "kitten" and can only answer MO-related product questions. If you want to learn more about world affairs or emotional topics, please refer to Google.

[Feature 3] Supports Private Link Connection

Private Link is a common way to access cloud databases in a production environment. Compared to public network connections, it's secure, reliable, and easy to configure. Its specific advantages include:

Private Network Communication*: Accessing endpoint services via Private Link ensures all access traffic is forwarded within the Alibaba Cloud intranet, avoiding public network exposure and potential security risks. Secure and Controllable: When accessing cloud services via Private Link, security group rules can be added to the Elastic Network Interface (ENI) used for service access in the VPC network, providing stronger security guarantees and control. Low Latency and High Quality: Access requests via Private Link are forwarded within the same availability zone, offering optimal latency. *Simple Management: Flexible cross-account and cross-VPC service access avoids complex routing and security configurations.

[Feature 4] Supports Chinese Page Interface

While supporting an English interface when we launched on AWS overseas was understandable, not supporting Chinese when we launched on Alibaba Cloud would be a significant oversight. Therefore, we've decided to fully localize the platform to Chinese. For domestic users, the Chinese interface will be displayed by default, though you can switch languages based on your preference. Every decision we make aims for you to use MO Intelligence more elegantly.

Improvements

  • Added a new"limit" status for instances, used to indicate situations where an instance has reached consumption limits, exhausted free CU, run out of free storage, or has insufficient account balance.
  • Optimized thequery editor's execution feedback time. By introducing a WebSocket mechanism, the query editor's execution feedback time has been significantly reduced.
  • Addedplatform free instance CU usage reminders. At the top of the instance management platform, we will provide real-time reminders of the CU usage of all free instances on the platform.
  • RefinedCU calculation rules. Public network egress traffic is now factored into the calculation of basic CU and SQL CU.
  • MatrixOne Intelligence now uniformly uses1024-based units for measuring memory and object storage, e.g., 1MB = 1MiB = 1024 KB.

Bug Fixes

  • Fixed an issue whereinstance storage value updates failed.
  • Fixedpaid instance creation failures.
  • Fixedbilling details search failures.
  • FixedLoad data failures in the query editor.
  • Fixed an issue where SQL still generatedpublic network traffic when using private link connections.
  • Fixed an issue where free-to-paid instances disappeared after being terminated and then recovered.
  • Fixeddatabase management platform login failures.

August 12, 2023

Features

[Feature 1] Official Support for Alibaba Cloud

In this version, we officially support the deployment of MO Intelligence on a domestic cloud provider –Alibaba Cloud. After verification tests with several customer scenarios, cloud instances have demonstrated high availability. MO Intelligence on Alibaba Cloud also providesfree Serverless instances for testing. We welcome everyone to apply for a trial.

[Feature 2] Supports Auto-Scaling Serverless Paid Instances

In this version, we introducedServerless paid instances with independent compute resources. Compared to free Serverless instances with shared compute resources, paid instances offer better performance and control, specifically:

  • MO Intelligence can detect the computing power demands of paid instances andautomatically scale compute resources up and down. This means your business can consistently experience high-performance SQL services without needing to manage resource changes.
  • When compute resources automatically scale down, existing connected sessions remain available, so you don't have to worry about service interruptions caused by compute node reduction.
  • If you've had a good experience with previous or current free instances and wish to expand your usage, there's no need to recreate paid instances and import data. We provide aone-click conversion button from free instances to paid instances for flexible use.
  • We've addedconsumption limits and service retention configurations for paid instances. You can set reasonable daily/monthly maximum consumption limits based on your business budget. Instances are billed hourly. When the consumption limit is reached, instance access will be stopped. Alternatively, you can configure the service retention feature to continue consuming at a lower rate after reaching the consumption limit, ensuring minimum business availability.
  • Added relevantmonitoring alerts: email alerts when paid instance consumption approaches daily/monthly limits, and email alerts when paid instance CU consumption exceeds 10M within the past 10 minutes. [Feature 3] Supports SQL Workbook Functionality

If you frequently use MO Intelligence's database management platform for database operations and maintenance, this feature is perfect for you. We now offeronline SQL text editing and saving functionality. After logging into the platform, you can access your historical SQL scripts online, eliminating the need to save and transfer them using other text files. The specific specifications for SQL Workbooks are as follows:

  • MO Intelligence provides each SQL User with up to100 SQL workbook records.
  • Each SQL workbook can store up to1M characters of text.
  • Each SQL workbook can store up to25 historical versions. If you accidentally make an incorrect modification to a workbook and want to undo it, or if you want to view a workbook from a few days ago, this feature fully meets your needs. [Feature 4] Query History SQL Aggregation

When businesses execute highly concurrent operations, a large number of SQL statements can be generated in a short time (e.g., 100,000 insert and select SQLs in 5 seconds). Previously, query history would record every single SQL statement. We believe such detailed recording is unnecessary because users typically care about statistical values over a period rather than the execution details of every statement. Overly detailed and numerous SQL entries can hinder observation and also consume significant computing resources, impacting performance under high concurrency. Therefore, we will nowaggregate SQL statements that meet certain rules at regular intervals. The specific rules are as follows:

  • Insert, Update, Delete, Execute, and Select SQL statements with an execution time ofless than 200ms will be included in the aggregation scope.
  • Aggregation will occurevery 5 seconds. Aggregated records will include the count of aggregated SQLs, total execution time, original SQL statement, etc., but will no longer record the execution plan of each individual SQL.
  • SQL executed from the SQL Editor in the database management platform willnot be aggregated.

Improvements

*MO Intelligence timezone display optimization. We've added a timezone display and switching function in the top right corner of the cloud platform. This helps you understand the timezone information for various times displayed on the platform. If you're Browse instances outside your region, you can switch to your local timezone for easier viewing.

  • Addedstorage limits for Serverless free instances. If an instance's storage size exceeds 5GB, SQL write operations for that instance will be prohibited, allowing only SELECT, SHOW, and DELETE operations. Once data is deleted and the size falls below this threshold, normal operations will resume immediately.
  • Improved thereal-time nature of CU statistics in Serverless instances (optimized from 80 minutes to less than 30 minutes).
  • Added aone-click load for TPC-H 10G sample data in Serverless instances.
  • Instance monitoring and database usage monitoring now providefiner-grained 30-minute and 1-hour interval statistics.
  • Added display of connection information forGo language to instances.
  • OptimizedMatrixOne Intelligence's machine configuration on Alibaba Cloud.
  • Optimized the display behavior wheninstance storage information retrieval fails.
  • Significantly improved theresponse performance of various cloud platform information (e.g., instance details, monitoring, query history).
  • Optimized the display oflong field information in the interface.
  • Optimized the display of the SQL Editor interface when switching page menus and returning.

Bug Fixes

  • Fixed an issue whereinstance storage size values were incorrect.
  • Fixed occasionalautomatic logout from the database management platform.
  • Fixed an issue wheretrial application emails for MO Intelligence were invalid.
  • Fixed an issue wherefailed SQL statements were not recorded in query history in some scenarios.
  • Fixedslow Serverless instance creation.
  • Fixed an issue where Serverless paid instances could still be accessed after reaching their consumption limit.
  • Fixedincorrect table statistics in the SQL Editor.
  • Fixedmissing sample data statistics.
  • Fixed an issue where terminated instances could not be recovered in some scenarios.
  • Fixed an issue where Serverless paid instances experiencedservice interruptions during high concurrency.

June 09, 2023

Features

Supports Alibaba Cloud (Beta)* Supports Paid Instances (Beta)*

  • Providesone-click quick creation of TPC-H databases using the publish-subscribe feature, saving import costs.
  • Addedtable details in the SQL Editor's data object panel, making it easier for users to write SQL.
  • AddedSQL-level CU viewing functionality, allowing users to view the CU consumption of each SQL statement.
  • AddedDatabase information for SQL execution in the SQL Editor's execution results.
  • AddedSQL execution result row count in query history.
  • Supportsresetting administrator passwords in instance details.
  • Differentiated SQL execution types in the user operation interface within the database management platform.

Improvements

  • Reduced thestatistical delay for instance storage, allowing instance storage size to be visible within one minute of instance creation.
  • Optimized the communication method between instance VPC and S3 from NAT to endpoint, saving costs and enhancing security.
  • Optimized theSQL Editor's interface interaction logic, enhancing usability.
  • Improved theSQL execution performance of import interactions, automatically adding parallel import mode.
  • EnhancedMO Intelligence API field validation, making cloud platform access more secure.
  • Optimized thesize and style of the SQL execution plan tree diagram, making it easier for users to view execution plans.
  • Optimized thepage layout for instance creation.
  • Optimized thesliding range of monitoring charts in the dashboard, ensuring that platform top information remains visible when scrolling through monitoring data.
  • Optimized thesliding range size when instance creation page information scrolls.
  • Corrected an issue where monitoring information had values even before instance creation; numerical values have also been added for monitoring data at other times.
  • Addedempty state display information for the SQL Editor results pane.
  • Optimized the scope of file permissions obtained when using Role ARN secure access for interactive LOAD operations.
  • Optimized theoperable range of instances in different states.
  • Changed theTitle display of the MO Intelligence webpage.
  • Optimized the interaction logic foradding local IP addresses in the IP whitelist feature.

Bug Fixes

  • Fixed an issue where entering % in the session id filter condition in query history returned a 500 error.
  • Fixed an issue where entering % in the SQL Text filter condition in query history returned incorrect results.
  • Fixed an issue where selecting a start time greater than the current time in query history returned a 500 error.
  • Fixed an issue where the '/*cloud_nonuser*/' tag was not removed in query history.
  • Fixed incorrect CU display values in query history.
  • Fixedquery result download failures.
  • Fixed an issue where terminating instances whose names started with a number via API failed.
  • Fixed an issue where paid instances remained in a "creating" state indefinitely.
  • Fixed an issue where the database management platform displayed overflow if the instance name was too long.
  • Fixed incorrect instance storage size.
  • Fixed an issue where CN pods were not released after a paid instance was terminated.
  • Fixed an issue where executing USE DATABASE in the SQL Editor would not complete.
  • Fixed an issue where creating a database in the SQL Editor resulted in an error.
  • Fixed incorrect creation time displayed for sample data.
  • Fixed an issue where the data objects were not updated after creating a database in the SQL Editor.
  • Fixed tpch q7 execution failures.
  • Fixed a missing hint that "day is less than month" when editing the spend limit.
  • Fixed an issue where non-integer values could not be entered for the spend limit.
  • Fixed incorrect monthly/daily consumption values in the spend limit.
  • Fixed an issue where the consumption value remained empty after enabling spend limit editing.
  • Fixed an issue where the instance status was incorrect after an overdrafted instance was terminated and then recovered.
  • Fixed an issue where the Alibaba Cloud icon was not updated in the database management platform.
  • Fixed an issue where the database management platform unexpectedly logged out during use.
  • Fixed an error when deleting a sample database in the database management platform.
  • Fixed an issue with failed Alibaba Cloud trial invitation.
  • Fixed incorrect database information in the database details function.
  • Fixedunstable instance states.