How to use the command line to do health check for redis

How can I use the command line to check if redis is working as expected?

Redis-cli basics

Using redis-cli to check the health of the database sounds simple, in fact, it is: simple redis-cli PING on the Redis host returns PONG . This doesn't look good, but it is true. The redis function returned to PONG is normal and healthy.

In more detail: This means that the data set is fully loaded and Redis is ready to connect. If something does not work properly, it will display an error message, such as "(error) LOADING Redis is loading the dataset in memory". This will return when loading the rdb/aof file or waiting for the copy to complete.

By using redis-cli, you can easily use info commands and subcommands to get more statistics and information.

Command list

To repeat the command, you can easily use two options for monitoring: redis - cli - r 5 - i 2 <command> will execute <command> 5 times with an interval of 2 seconds.

Using these commands, you can easily integrate redis into your monitoring. Just add the redis-cli command as a custom parameter, for example to Zabbix or Nagios. You can also use prometheus exporter to get metrics from redis.

Integrated delay monitoring, reporting, slow log

Maybe some of you will ask yourself why we need monitoring in the first place, because redis is very fast. The answer is: you are right, redis is designed to process a large number of queries per second for each instance. Unfortunately, in most applications, there are strict requirements for average response time and worst-case latency.

Nevertheless, because redis has a large number of commands, many and complicated, they are also divided into three, six or nine levels. Some commands run very fast and run in constant or logarithmic time, while other commands are slower and may cause delay peaks. Other reasons for adopting this monitoring method are related to the interaction of the operating system (such as disk persistence) and the single-threaded architecture of redis.

Since version 2.8.13, redis has introduced latency monitoring, which can help solve possible latency issues. The following mechanisms are part of this framework:

• Use delay hooks to sample the paths of different delay-sensitive codes
• Time series recording of delay peaks divided by other events
• The reporting engine obtains raw data from the time series
• The analysis engine provides human-readable reports and prompts based on the measurements

How to enable delay monitoring?

When redis is running, latency monitoring can be easily enabled CONFIG SET latency-monitor-threshold 100 The mantissa defines the time in milliseconds. In our example, every event that takes more than 100ms will be recorded as a delay peak.

Use the LATENCY command to report information

The user interface of the delay monitoring subsystem is the LATENCY command. Like many other Redis commands, LATENCY accepts subcommands that modify its behavior. These subcommands are:

• LATENCY LATEST-Returns the latest delayed sample of all events.
• LATENCY HISTORY-Returns the delayed time series of a given event.
• LATENCY RESET-Reset the delayed time series data of one or more events.
• LATENCY GRAPH-presents an ASCII art graph of delayed samples of events.
• LATENCY DOCTOR-Reply to human-readable delay analysis report.

For details, please refer to the documentation page of each subcommand

What should I do when redis encounters latency spikes?

The next step may be to enable slow logging. The configuration command is config set slowlog-log-slower-than 1000 , you can record every event that lasts longer and exceeds 1000ms.