Logs Analytics

This is the guide to understand the logs a SSDB server generates.

Generally, I recommend you set logger.level to debug.

Request Handling

2014-06-18 11:01:40.335 [DEBUG] serv.cpp(395): w:0.393,p:5.356, req: set a 1, resp: ok 1

• w:0.393 request wait time, in ms
• p:5.356 request process time, in ms
• req:... the request
• resp:... the response

Find out Slow Queries

The Shell commands to find out slow queries:

tail -f log.txt | grep -w resp | grep '[wp]:[1-9][0-9]\{0,\}\.'
# or
cat log.txt | grep -w resp | grep '[wp]:[1-9][0-9]\{0,\}\.'

These commands find out requests with wait time or process time greater than 1ms.

Find out requests slower than 10ms:

cat log.txt | grep -w resp | grep '[wp]:[1-9][0-9]\{1,\}\.'

Find out requests slower than 100ms:

cat log.txt | grep -w resp | grep '[wp]:[1-9][0-9]\{2,\}\.'

SSDB is Working

ssdb-server print out log like these every 5 minutes:

2014-06-18 11:18:03.600 [INFO ] ssdb-server.cpp(215): ssdb working, links: 0
2014-06-18 11:23:03.631 [INFO ] ssdb-server.cpp(215): ssdb working, links: 0

• links: 0 current connections