un1ko85

#
# CORS header support
#
# One way to use this is by placing it into a file called "cors_support"
# under your Nginx configuration directory and placing the following
# statement inside your **location** block(s):
#
#   include cors_support;
#
# As of Nginx 1.7.5, add_header supports an "always" parameter which

un1ko85

Docker compose as a systemd unit

Create file /etc/systemd/system/[email protected]

[Unit]
Description=%i service with docker compose
Requires=docker.service
After=docker.service

un1ko85

version: "2.1"
services:
  apm-server:
    image: docker.elastic.co/apm/apm-server:${STACK_VERSION:-6.5.0}
    ports:
      - "127.0.0.1:${APM_SERVER_PORT:-8200}:8200"
      - "127.0.0.1:${APM_SERVER_MONITOR_PORT:-6060}:6060"
    command: >
      apm-server -e
        -E apm-server.rum.enabled=true

un1ko85

Intro

This is an example configuration to have nginx output JSON logs to make it easier for Logstash processing. I was trying to get nginx > Filebeat > Logstash > ES working and it wasn't until I connected Filebeat directly to Elasticsearch that I saw the expected data. Google led me to ingest-convert.sh and I realized filebeat setup works for Filebeat > ES but not Filebeat > Logstash > ES. This is because Logstash does not use ingest pipelines by default. You have to enable them in the elasticsearch output block.

Having nginx log JSON in the format required for Elasticsearch means there's very little processing (i.e. grok) to be done in Logstash. nginx can only output JSON for access logs; the error_log format cannot be changed.

Extra fields are output and not used by the Kibana dashboards. I included them in case they might be useful. Since they are not declared in the filebeat setup, their default is "string" when yo

un1ko85

import plotly.graph_objects as go
import pandas as pd

def serp_heatmap(df, num_domains=10, select_domain=None):
    df = df.rename(columns={'domain': 'displayLink',
                            'searchTerms': 'keyword'})
    top_domains = df['displayLink'].value_counts()[:num_domains].index.tolist()
    top_domains = df['displayLink'].value_counts()[:num_domains].index.tolist()
    top_df = df[df['displayLink'].isin(top_domains) & df['displayLink'].ne('')]

un1ko85

CREATE OR REPLACE FUNCTION pnorm(z double precision) RETURNS double precision AS $$
  SELECT CASE 
  	WHEN $1 >= 0 THEN 1 - POWER(((((((0.000005383*$1+0.0000488906)*$1+0.0000380036)*$1+0.0032776263)*$1+0.0211410061)*$1+0.049867347)*$1+1),-16)/2
  	ELSE 1 - pnorm(-$1)
  END;
$$ LANGUAGE SQL IMMUTABLE STRICT;