This is an analysis tool for funnel visualization with log from Elasticsearch. Even though we have Kibana can display log very well, it can't fit our goal that to analyze series log with context.
In order to run elastic_funnel, some works have to prepare.
- You have to run your Elasticsearch first
- Set the essential config file with your Elasticsearch environment
Pandas is one of core requirements and it could take a few minutes to complete.
pip install elastic_funnel
Add the argument file ~/.elastic_funnel to your home path, with
[elastic]
host = 127.0.0.1
port = 9200
index = user-behavior-*
query = action:state_change AND sessionid:*
[fields]
timestamp = @timestamp
stage = state
identity = sessionid
elastic_funnel --stages landingpage login searchpage --start 2016-03-25T00:00:00
You could get a visualization with ASCII characters looks like follows, the percentage means trend from one stage to the next one.
Funnel: landingpage --> login --> searchpage
###############################################################################
██████████████████████████████████████████████████ 27 100.0% landingpage
██████████████ 8 29.6% login
█ 1 12.5% searchpage
usage: elastic_funnel [-h] [--host HOST] [--port PORT] [--index INDEX]
--stages STAGES [STAGES ...] [--start START] [--end END]
[--add_query ADD_QUERY]
optional arguments:
-h, --help show this help message and exit
--host HOST Host of Elasticsearch
--port PORT Port of Elasticsearch
--index INDEX Index name of Elasticsearch, e.g., user-behavior-log-*
--stages STAGES [STAGES ...]
Set a path of stages , e.g., index explore user
explore
--start START Start time of log, e.g., 2016-03-24T00:00:00
--end END End time of log, e.g., 2016-03-28T00:00:00
--add_query ADD_QUERY
Additional query using syntax of Lucene, https://lucen
e.apache.org/core/2_9_4/queryparsersyntax.html. You
can narrow you search target by syntax, e.g.,
country:US
docker pull yuecen/elastic_funnel
docker run -it --rm -v ~/.elastic_funnel:/root/.elastic_funnel:ro yuecen/elastic_funnel elastic_funnel --stages landingpage login searchpage
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent