Usage:
redis-look-monitor.py
--host,-h <host> (default: localhost)
--port,-p <port> (default: 6379)
--estimate-throughput,-e (default: false)
--estimate-throughput-limit,-l <num> (default: 1000)
--input-file <filename> (default: None)
--summary-number,-n <num> (default: 5)
By default, the tool will connect to Redis and run MONITOR until Ctrl-C is pressed. It will then output the top operations for various dimensions. Note: MONITOR will negatively affect overall throughput, take care in production.
If --estimate-throughput is set, the tool will take the output from MONITOR and query DEBUG OBJECT for the top <limit> keys to get the serialized length. It will take the length returned and multiply it by the number of times accessed to get a rough estimate of the throughput required to service a key. Currently, the tool is not read/write aware, and it does not handle MGET properly which will affect the accuracy of the estimate.
Example
pip install requirements.txt
./redis-look-monitor.py -e
Connecting...
Issuing MONITOR...
Reading commands... 982
^C 982 commands in 4.09 seconds (240.37 cmd/s) across 743 unique keys
* top by key
count avg/s % key
32 7.83 3.3 _lock
3 0.73 0.3 __raw__327
3 0.73 0.3 __raw__553
3 0.73 0.3 __raw__334
3 0.73 0.3 __raw__113
* top by command
count avg/s % command
636 155.68 64.8 SETEX
104 25.46 10.6 LTRIM
104 25.46 10.6 EXPIRE
104 25.46 10.6 LPUSH
32 7.83 3.3 EXISTS
* top by command and key
count avg/s % command and key
32 7.83 3.3 EXISTS last_113
1 0.24 0.1 SETEX last_999
1 0.24 0.1 SETEX last_596
1 0.24 0.1 LTRIM __raw__596
1 0.24 0.1 SETEX __raw__621
Estimating throughput requirements for top 1000 keys...
* top by est. throughput
est. bytes count throughput throughput/s key
1.0M 3 3.0M 0.8K __raw__222
3.3K 3 10.0K 2.4K __raw__103
3.3K 3 10.0K 2.4K __raw__990
3.1K 3 9.4K 2.3K __raw__327
3.1K 3 9.4K 2.3K __raw__553
TODO:
- create module and upload to PyPI
- fix various code TODOs