1. Background
The HUAWEI CLOUD GaussDB (for Influx) cluster version database has built features such as high reliability, high availability, high performance, high scalability, and O&M-free, so that customers do not have to worry about database-level issues, but rather focus on their own application development and design. Become the business first choice of many customers.
Compared with traditional databases such as MySQL, "new" related databases are quite different in terms of design and use. The data model and usage methods are relatively unfamiliar to most developers, and more instances are needed for testing and debugging. In non-production environments, such as development and testing scenarios, there is no demand for performance and reliability, and cost becomes an annoying consideration; if you choose your own construction environment for development and testing, although the cost problem is solved, maintenance is troublesome , The operation is complicated, and problems cannot be dealt with in time, which also greatly affects the efficiency of development and testing.
Therefore, GaussDB (for Influx) has launched a stand-alone version solution for sensitive usage scenarios with low performance and reliability requirements. This solution is used in scenarios such as development and testing. It can not only enjoy the convenience brought by service, but also significantly reduce the cost of use.
2. Maximum cost of GaussDB (for Influx) stand-alone version
When using databases, many customers feel that they can purchase ECS to build databases without considering high availability, performance, and scalability testing and development scenarios. Although this is troublesome, it can reduce the cost of use. After all, cloud databases are also based on Established by ECS, "bread" must be more expensive than "flour".
The GaussDB (for Influx) team combined the practical experience of many parties, comprehensively considered various factors such as cost, stability, and transportability, and chose a 1:2 ultra-high upgrade specification for the stand-alone version solution, and supported 4 different specifications at the same time to meet For the requirements of different scenarios, the following are the types of specifications supported by the stand-alone version and the corresponding fees:
| Specification |
Stand-alone version (yuan/month) |
Self-built ECS (yuan/month) |
| 4U8G |
486.9 |
Chapter 489 |
| 8U16G |
973.8 |
950 |
| 16U32G |
1947.6 |
1712 |
| 32U64G |
3895.2 |
3396 |
Note: The cost comparison is only for computing resources, and storage resources are purchased according to the amount of data
It can be seen that the price of the stand-alone version of GaussDB (for Influx) is basically the same as that of self-built. Under the 4U8G specification, it is even cheaper than buying ECS for self-built. The stand-alone version of GaussDB (for Influx) not only achieves the same price of "flour" and "bread", but also provides many service capabilities that self-built does not have.
3. Advantages of GaussDB (for Influx) stand-alone version
3.1 Cloud services are out-of-the-box, free of operation and maintenance
Although the database is basic software, its architecture also requires a series of basic services as support. After purchasing ECS on the cloud, customers usually need to select the appropriate database version and modify the appropriate configuration parameters according to the current ECS specifications to achieve high availability of basic configuration parameters. After completing the above work, it is necessary to build a monitoring system to ensure basic operation and maintenance. In contrast, customers who use cloud databases do not need to worry about the above-mentioned complicated work at all, and can concentrate on the realization and optimization of their own business, as shown in the following figure:
Especially when creating and deleting environments in batches and intermittently, the manual operations described above can be cumbersome, running, and error-prone. After the database is created, subsequent operation and maintenance is also a test of experience. If it is an unfamiliar database, a simple problem may hinder development and testing, and may eventually affect the overall details of the project.
In contrast, the operation of cloud database is very simple and convenient. You only need to simply click on the page to purchase and create, select the corresponding storage size, and fill in the password, which greatly improves the operation efficiency. The layout is that when encountering problems, there is an experienced operation and maintenance team behind the cloud database to ensure that common problems can be recovered within minutes.
In addition to the convenience of use and deployment, the cloud database kernel has made many normative restrictions on unreasonable business use, so that customers can find problems in the business development and testing stages, and optimize and adjust them in time to ensure that the official business is stable. .
3.2 Limit unreasonable queries
GaussDB (for Influx) has made many restrictions and specifications at the database kernel level. Unreasonable usage is not allowed, and suggestions for optimization are given. Following these specifications can effectively improve the stability of business operations.
- Disable query execution without a time range
Data has a large number of time attributes, and the internal organization of the database also strictly follows the time attributes, which is a big difference from other databases. If the query does not include a time range, the database will query all the data, which has a high probability of causing OOM problems in the database and affecting business stability.
- Prohibit the execution of select * from statement query
Most real business scenarios need to specify specific custom query fields, but select * from is used for query due to various reasons such as development. This way of writing has two potential problems: 1. The result may be uncertain. GaussDB (for Influx) is in schemaless mode, and tags and fields can be added and reduced at any time, so that the results returned by select* may appear as the business changes Inconsistency may easily lead to business exceptions when parsing the returned results; 2 select* returns the values of all fields, but in fact only a few fields are useful, so a large amount of data transmission is invalid, wasting network bandwidth and wasting Computing, memory and other resources. Therefore, it is not recommended to use this method of query in the development specification, and this query is intercepted in the kernel.
- View slow log capability
The slow log capability of the cloud service enables timely discovery of slow queries during the business development phase and timely optimization and adjustment.
- The number of timelines and fields involved in each limit query
If a query involves too many timelines or fields, it means that the query may query too much data, and the database will load these data into the memory, which will easily lead to OOM and business stability. However, too many timelines or fields are involved, and most of the following situations are caused by the unreasonable query statements we wrote, which need further optimization.
4. Summary
GaussDB (for Influx) is a cloud-native database based on the separation of computing and storage architecture, compatible with the InfluxDB ecology, and the first stand-alone version mode. While retaining the advantages of cloud databases, it is basically the same price as self-built ECS, and meets business requirements for availability, The demand for bright spots in scenarios with low reliability requirements.
Click to follow and learn about Huawei Cloud's fresh technologies for the first time~