Growing with ElasticSearch
The Journey
- 2014
- Postgres full text search - only for admins - 5K users - 2 devs
- 2015
- ElasticSearch (v1.4) powered search - for 25 K users - 3 devs
- 2016
- Recommendations to users, search algorithms
- Log monitoring for devs with Filebeat + Logstash
- 1 lac users - 7 devs
- 2017
- Kibana based analytics - Filebeat + Logstash
- Powerful searches to give the users what they are looking for
- 2018 - ElasticSearch (v5.5) for recommendations and powerful searches, log
monitoring solution for dev team, business analytics for higher management
- 1M users - 22M docs - 5 devs
- 2018 Roadmap
- Machine learning and anomaly detection with v6.2
- Monitoring ES cluster and securing within ElasticStack
The Journey Contd...
Search for a search engine
- Postgres full text search wasn’t enough
- Why ElasticSearch and not Solr or Sphinx ?
The goodness of ElasticSearch
- Flexible Query DSL and REST APIs - ES queries are artistic creations :)
- Scaling comes for free
- Data aggregations are way to analytics
- The stack around - suitable for data pipelines and analytics
- Useful plugins
- Up-to-date documentation and client libraries even with a very fast paced
release cycle
- Helpful community
Log monitoring: (would-be-)paid solution to Elastic based
Kibana - best pal to visualize
- Any data - time series, geo data etc
- Any chart - histograms, line graphs, pie charts, sunbursts, and more
- Usability - non techies also can use it
- Export and share the dashboards
- [share screenshots]
Architecture
[Architecture Diagram of our data pipelines involving Elastic Stack, few micro
services and dozens of servers]
Performance tuning
- very slow queries and how we solved them
- Things to look into when the performance is not optimal
Monitoring and Securing the cluster
- how we deal with the cluster
- How v6.2 is going to change this landscape
Questions

Growing with elastic search

  • 1.
  • 2.
    The Journey - 2014 -Postgres full text search - only for admins - 5K users - 2 devs - 2015 - ElasticSearch (v1.4) powered search - for 25 K users - 3 devs - 2016 - Recommendations to users, search algorithms - Log monitoring for devs with Filebeat + Logstash - 1 lac users - 7 devs - 2017 - Kibana based analytics - Filebeat + Logstash - Powerful searches to give the users what they are looking for
  • 3.
    - 2018 -ElasticSearch (v5.5) for recommendations and powerful searches, log monitoring solution for dev team, business analytics for higher management - 1M users - 22M docs - 5 devs - 2018 Roadmap - Machine learning and anomaly detection with v6.2 - Monitoring ES cluster and securing within ElasticStack The Journey Contd...
  • 4.
    Search for asearch engine - Postgres full text search wasn’t enough - Why ElasticSearch and not Solr or Sphinx ?
  • 5.
    The goodness ofElasticSearch - Flexible Query DSL and REST APIs - ES queries are artistic creations :) - Scaling comes for free - Data aggregations are way to analytics - The stack around - suitable for data pipelines and analytics - Useful plugins - Up-to-date documentation and client libraries even with a very fast paced release cycle - Helpful community
  • 6.
    Log monitoring: (would-be-)paidsolution to Elastic based
  • 7.
    Kibana - bestpal to visualize - Any data - time series, geo data etc - Any chart - histograms, line graphs, pie charts, sunbursts, and more - Usability - non techies also can use it - Export and share the dashboards - [share screenshots]
  • 8.
    Architecture [Architecture Diagram ofour data pipelines involving Elastic Stack, few micro services and dozens of servers]
  • 9.
    Performance tuning - veryslow queries and how we solved them - Things to look into when the performance is not optimal
  • 10.
    Monitoring and Securingthe cluster - how we deal with the cluster - How v6.2 is going to change this landscape
  • 11.