Neo4j Aura Query Logs Filtering for Slow Queries On Mac / Linux

First you will need to have a log file.  Please see this article, How To Download Query Log from Aura Instances, if you need information on how to get a query log.  You can also view the documentation on logs for Aura.

Second, this code was tested on mac / linux, and uses the basic commands of grep and awk.  However, no expertise is needed to run these commands.  All you need is a downloaded log file, and a shell/terminal to execute the below commands.  Assumes you are in the directory of the file called query.log, and will output a new file in that directory.  Please see the examples below.

Example 1, filter out queries that take more than 1 second (1,000ms) from the file (query.log) in the current directory and into a file called slow_queries_extracted_1sec.log:

grep -i '"elapsedTimeMs"\|"query":' ./query.log | awk '/elapsedTimeMs/ {if ((substr($2,0,length($2)-1)+0)>1000) {print;getline;print;}}' > slow_queries_extracted_1sec.log 

Example 2, filter out queries that take more than 30 seconds (30,000ms) from the file (query.log) and into a file called slow_queries_extracted_30sec.log:

grep -i '"elapsedTimeMs"\|"query":' ./query.log | awk '/elapsedTimeMs/ {if ((substr($2,0,length($2)-1)+0)>30000) {print;getline;print;}}' > slow_queries_extracted_30sec.log 

Example 3, filter out queries that take more than 1 minute (60,000ms) from the file (query.log) and into a file called slow_queries_extracted_1min.log:

grep -i '"elapsedTimeMs"\|"query":' ./query.log | awk '/elapsedTimeMs/ {if ((substr($2,0,length($2)-1)+0)>60000) {print;getline;print;}}' > slow_queries_extracted_1min.log 

Example 4, filter out queries that take more than 2 minutes (120,000ms) from the file (query.log) and into a file called slow_queries_extracted_2min.log:

grep -i '"elapsedTimeMs"\|"query":' ./query.log | awk '/elapsedTimeMs/ {if ((substr($2,0,length($2)-1)+0)>120000) {print;getline;print;}}' > slow_queries_extracted_2min.log