Tair is a high-performance database service. This topic describes the development and O&M standards that you can follow to design a more efficient business system and better utilize Tair. The standards are developed by Alibaba Cloud based on years of O&M experience and are applicable to the following scenarios: business deployment, key design, SDK usage, command usage, and O&M management.

Understand the performance limits of Tair

Figure 1. Performance limits of Tair
Performance limits of Tair
Resource type Description
Computing resources Wildcard characters, concurrent Lua scripts, one-to-many PubSub commands, and hotkeys consume a large amount of computing resources. For cluster instances, these items can also cause skewed requests and underutilized data shards.
Storage resources Streaming jobs and large keys consume a large amount of storage resources. For cluster instances, these items can also cause data skew and underutilized data shards.
Network resources Database-wide scans (by running the KEYS command) and range queries of big values and large keys (by running the HGETALL command) consume a large amount of network resources and often cause thread congestion.
Important The high-concurrency capability of Tair does not efficiently improve access performance as expected but does affect the overall performance of Tair. For example, the storage of big values in Tair does not improve access performance.

For cluster instances, large keys, hotkeys, and great values may cause skewed storage or skewed requests.In a production environment, you must prevent reaching the performance limits of Tair. The following tables describe the business deployment, key design, SDK usage, command usage, and O&M management standards for Tair. These standards help you design a more efficient business system and better utilize the capabilities of Tair.

Business deployment standards

Importance Standard Description
★★★★★

Determine whether the scenario is high-speed cache or in-memory database.

  • High-speed cache: To reduce overheads and prevent strong dependence on the data in a cache, we recommend that you disable append-only file (AOF) in cache-only scenarios because the data may be evicted. For more information about AOF, see Disable AOF persistence. For example, after a Tair database is full, the data eviction policy is triggered to reclaim space for writing new data. The latency increases with the amount of data that is written.
    Important To use the data flashback feature, you must enable AOF. For more information, see Use data flashback to restore data by point in time.
  • In-memory databases: We recommend that you choose Tair persistent memory-optimized instances that offer command-level persistence. In addition, you can monitor memory usage by configuring alerts in the databases. For more information, see Alerting settings.
★★★★★ Deploy your business close to Tair instances. For example, you can deploy your business in an Elastic Compute Service (ECS) instance that resides in the same virtual private cloud (VPC) as your Tair instances. Tair is a high-performance database service. However, if you deploy your business server far from Tair instances and the business server and instances are connected over the Internet, the performance of Tair is greatly reduced due to network latency.
Note For cross-region deployment, you can use the geo-replication capability of Global Distributed Cache for Tair which is also known as Global Replica to implement geo-disaster recovery or active geo-redundancy, reduce network latency, and simplify business design. For more information, see Overview.
★★★★☆ Create a Tair instance for each service. Do not use a Tair for different services. For example, do not use a Tair instance for both high-speed cache and in-memory database services. Otherwise, the eviction policies, slow queries, and FLUSHDB command of one service affect other services.
★★★★☆ Configure appropriate eviction policies to evict expired keys. The default eviction policy for expired keys in Tair is volatile-lru.For more information about eviction policies, see Supported parameters.
★★★☆☆ Manage stress testing data and duration. Tair does not delete stress testing data. To prevent impacts on your business, you must manage stress testing data and duration by yourself.

Key design standards

Importance Standard Description
★★★★★ Configure key values to an appropriate size. We recommend that you configure key values to a size smaller than 10 KB. Excessively large values can cause data skew, hotkeys, high bandwidth, or high CPU utilization. You can prevent these issues from the beginning by making sure that key values are of proper size.
★★★★★ Configure proper key names that have proper length.
  • Key names:
    • Use readable strings as key names. If you want to combine a database name, table name, and field name into a key name, we recommend that you use colons (:) to separate them. Example: project:user:001.
    • Shorten key names without compromising their ability to describe your business. For example, username can be shortened to u.
    • In Tair, braces {} are recognized as hash tags. In this case, if you use cluster instances, you must correctly use braces in key names to prevent data skew.For more information, see Keys hash tags.
      Note For a cluster instance, if you want to manage multiple keys by running a command such as the RENAME command and do not use hash tags to ensure that the keys reside in the same data shard, the command cannot be run.
  • Length: We recommend that you configure key names to be no more than 128 bytes in length. The shorter, the better.
★★★★★ For complex data structures that support sub-keys, you must avoid including excessive sub-keys in one key. We recommend that you include less than 1,000 sub-keys in a key.
Note Common complex data structures include hashes, sets, Zsets, GEO structures, streams, and structures that are provided only by Tair, such as TairHash, TairBloom, and TairGIS.
The time complexity of specific commands, such as HGETALL, is directly related to the number of sub-keys. Excessive sub-keys increase the time complexity of a command. If you frequently run commands whose time complexity is O(N) or higher, many issues occur, such as slow queries, data skew, and hotkeys.
★★★★☆ Use the serialization method to convert values into readable structures. The bytecode of a programming language may change when the version of the language changes. If you store naked objects (such as Java objects and C# objects) in Tair instances, the software stack may be difficult to upgrade. We recommend that you use the serialization method to convert values into readable structures.

SDK usage standards

Importance Standard Description
★★★★★ Use JedisPool or JedisCluster clients to connect to Tair instances.
Note We recommend that you use TairJedis clients to connect to DRAM-based instances, because TairJedis clients support the encapsulation of new data structures. For more information, see TairJedis client.
If you use a single connection, the client cannot automatically reconnect to Tair instances after a connection times out. For more information about how to use JedisPool clients to connect to Tair instances, see Jedis client, JedisPool optimization, and JedisCluster.
★★★★☆ Do not use Lettuce clients. Lettuce clients do not automatically reconnect to Tair instances after multiple requests time out. If failures occur in a Tair instance and cause switchovers on proxy nodes or data shards, connections may time out and Lettuce clients cannot reconnect to the Tair instance. To prevent such risks, we recommend that you use a Jedis client to connect to Tair instances. For more information, see Jedis client.
★★★★☆ Design proper fault tolerance mechanisms for your clients. Network fluctuations and high usage of resources may cause connection timeouts or slow queries. To prevent these risks, you must design proper fault tolerance mechanisms for your clients.
★★★★☆ Set longer retry intervals for your clients. If retry intervals are shorter than required (example: less than 200 milliseconds), a large number of retries may occur in a short period of time. This tends to result in a service avalanche. For more information, see Retry mechanisms for Tair clients.

Command usage standards

Importance Standard Description
★★★★★ Avoid range queries, such as those by running the KEYS * command. Instead, use multiple point queries or run the SCAN command to reduce latency. Range queries may cause service interruptions, slow queries, or congestion.
★★★★★ Use extended data structures to perform complex operations. For more information, see Extended data structures of Tair. Do not use Lua scripts. Lua scripts consume a large amount of computing and memory resources and do not support multi-threading acceleration. Overly complex or improper Lua scripts may result in the exhaustion of resources.
★★★★☆ Use pipelines to reduce the round-trip time (RTT) of data. If you want to send multiple commands to a server and your client does not depend on each response from the server, you can use a pipeline to send the commands at a time. Take note of the following items when you use pipelines:
  • A client that uses pipelines exclusively connects to a server. We recommend that you establish a dedicated connection for pipeline operations to separate them from regular operations.
  • Each pipeline must contain a proper number of commands. We recommend that you use each pipeline to send no more than 100 commands.
★★★★☆ Use transaction commands. For more information, see Limits on commands supported by Tair. When you use transaction commands, take note of the following limits:
  • Transactions cannot be rolled back.
  • If you want to run transaction commands on cluster instances, use hash tags to ensure that the keys to be managed are distributed to the same hash slot. You must also prevent skewed storage that may be caused by hash tags.
  • Do not encapsulate transaction commands in Lua scripts, because the compilation and loading of these commands consume a large amount of computing resources.
★★★★☆ Do not use the Pub and Sub command group to perform a large number of message distribution tasks. For more information, see Limits on commands supported by Tair. The Pub and Sub command group does not support data persistence or acknowledge mechanisms that ensure data reliability. We recommend that you do not use Pub or Sub commands to perform a large number of message distribution tasks. For example, if you use these commands to distribute a message larger than 1 KB in size to more than 100 subscriber clients, server resources may be exhausted and subscriber clients may not receive the message.
Note To improve performance and balance, Tair is optimized for Pub and Sub commands. In cluster instances, proxy nodes calculate the hash values of commands based on channel names and allocate commands to corresponding data nodes.

O&M management standards

Importance Standard Description
★★★★★ Understand the impacts of different instance management operations. Configuration changes or restarts affect the state of a Tair instance. For example, transient connections may occur for the instance. Before you perform the preceding operations, make sure that you understand the impacts. For more information, see Instance states and impacts.
★★★★★ Verify the error handling capabilities or disaster recovery logic of a client. Tair can monitor the health status of nodes. If a master node in an instance becomes unavailable, Tair automatically triggers a master-replica switchover. The roles of master and replica nodes are switched over to ensure the high availability of the instance. Before a client is officially released, we recommend that you manually trigger the master-replica switchover. This can help you verify the error handling capabilities or disaster recovery logic of the client. For more information, see Manually switch workloads from a master node to a replica node.
★★★★★ Disable time-consuming or high-risk commands. In a production environment, abuse of commands may cause problems. For example, the FLUSHALL command can delete all data. The KEYS command may cause network congestion. To improve the stability and efficiency of services, you can disable these commands to minimize risks. For more information, see Disable high-risk commands.
★★★★☆ Handle pending events in a timely manner. To enhance user experience and provide improved service performance and stability, Alibaba Cloud occasionally generates pending events to upgrade the hardware and software of specific servers or replace network facilities. For example, a pending event is generated when the minor version of databases needs to be updated. After you receive an event notification from Alibaba Cloud, you can check the impacts of the event and change the scheduled time of the event to meet your business requirements. For more information, see View and manage pending events.
★★★★☆ Configure alerts for core metrics and better monitor the status of your instances. Configure alerts for core metrics such as CPU utilization, memory usage, and bandwidth usage to monitor the status of your instances in real time. For more information, see Alerting settings.
★★★★☆ Use O&M features provided by Tair to check the status of instances or troubleshoot resource usage exceptions on a regular basis.
  • Use slow logs to troubleshoot timeout issues: Slow logs help you locate slow queries and the IP addresses of the clients that send the query requests. Slow logs provide a reliable basis for addressing timeout issues.
  • View monitoring data: Tair supports a variety of performance metrics. These metrics allow you to gain insights into the status of Tair instances and troubleshoot issues.
  • Create a diagnostic report: Diagnostic reports help you evaluate the status of Tair instances, such as performance level, skewed requests, and slow logs. Diagnostic reports also help you identify exceptions on Tair instances.
  • Offline key analysis: You can use the offline key analysis feature to identify large keys of Tair instances. You can also learn the memory usage, distribution, and expiration time of large keys.
  • Use the real-time key statistics feature: The real-time key statistics feature helps you identify hotkeys of Tair instances and allows you to further optimize your databases.
★★★☆☆ Enable the audit log feature and evaluate audit logs. After you enable the audit log feature, the audit statistics about write operations are recorded. Tair also allows you to query, analyze online, and export audit logs. These features help you monitor the security and performance of your Tair instances. For more information, see Enable the audit log feature.
Important After you enable the audit log feature, the performance of Tair instances may degrade by 5% to 15%. The actual performance degradation varies based on the number of write operations or audit operations. If your business expects a large number of write operations, we recommend that you enable the audit log feature only when you perform O&M operations, such as troubleshooting. This helps you prevent performance degradation.