Infrastructure for elastic

[Sonea]

Hello Community,

at the moment, I would like to inform myself a bit more.
I did some research already but I didn't got a satisfying answer.
I would like to know on which infrastructure elastic can run?
I know about cloud.
But more interesting, what about SaaS and especially on premise, so local?
Is there a combination between these possible?

Thank you very much for your support and help!

• Sonea

[shanec (Shane Connelly)]

You can do all the things!

• Elastic has a hosted SaaS platform called Elastic Cloud where we'll host and manage Elasticsearch and Kibana for you.
• Already have an AWS bill and want Elastic Cloud to be added straight to it? We have marketplace integration for that.
• Want to deploy and manage your own Elasticsearch on AWS? You can do that too.
• Prefer GCP? You can launch that with 1 click and we're working to get our SaaS on GCP as well
• Prefer Azure? Yeah, you can do that too.
• Want to download yourself and deploy/run on-prem? Sure thing. Download all the components with our quick-start guide. We also have a Puppet, Chef, and Ansible bits if you have any of those as a preferred deployment/configuration management tool
• Use Docker? We make official images.
• Want to do your own private Elasticsearch cloud? Yep, we have you covered.

[Sonea]

Thank you very much! This helps me out a lot!

In connection to this, how flexible are the solutions of elastic? Is it possible to customize or is it mostly fix?

[shanec (Shane Connelly)]

I think this question is too abstract to be able to really answer well. I think there are 3 ways to go about getting started answering, but ultimately it's going to come down to the third.

First way to answer I can think of is: Elasticsearch, Logstash, Beats, and Kibana are all highly configurable and Elasticsearch has extremely rich RESTful APIs that allow you to build complex applications on top of it. They're also all pluggable or provide libraries to build from. For example, there are a variety of Elasticsearch plugins, Logstash input, filter, and output plugins, Kibana plugins, and Beats has a library that people have built a variety of community Beats from. But due to the richness of Elasticsearch's APIs, a lot of times users don't need to actually plug/extend any of the core APIs. Also, you can pick and choose which of the components you use. Some people just use Elasticsearch and build their own data loaders / apps / visualizations that show data on top. Others use just Elasticsearch+Kibana and have their own or 3rd party data shippers. Others just use Elasticsearch+Logstash or Elasticsearch+Beats or Elasticsearch+Beats+Logstash.

Second way may be by talking about the variety of example use cases. A small selection of them is outlined here.

• Elasticsearch is used for full-text search a lot of places. For example, it's what powers Wikipedia's search and you can find tons of examples of people using Elasticsearch for this type of use case.
• Because it's very good at full-text search (and other types of search), it can be used for Enterprise search. There are a variety of companies which build connectors that push enterprise data into Elasticsearch and/or provide an enterprise search UI on top (as well as open source frameworks to do similar).
• It's also used by a number of organizations for geo search and analytics (e.g. ArcGIS, Uber, USGS and others), as there are rich geo search and analytics functions in Elasticsearch and Kibana.
• Probably the most well known use case for the Elastic Stack is operational log analytics. You can find thousands of users doing this, often referred to by this use case's older moniker "ELK," which stands for "Elasticsearch, Logstash, and Kibana."
• The same log analytics platform is also used by organizations like USAA to power security analytics and threat hunting.
• It's often used as a BI / business analytics tool, partially because Kibana is a pretty rich tool that can sit on top of Elasticsearch and visualize business data, partially because Elasticsearch has built-in columnar storage in the form of doc_values, and partially because it's nice to have your logs feed directly into your BI tool, since things like KPIs often can be deduced from logs anyway. Again, you can find a variety of users doing exactly this, like Yodle and The Guardian.
• It can be used for metrics data, like NASA has done with shipping telemetry data back from the Mars rover. We're seeing an uptick in metrics data use cases, for a few reasons. First, because we have Metricbeat, which provides an easy way to ship metrics data. Second, because Elasticsearch came out with specialized structure in 5.0 for dealing with numeric data more efficiently. Third, because Kibana is a great visualization tool for it, especially with Timelion and Time Series Visual Builder (just released in 5.4) in Kibana. And fourth, because a lot of times it just makes sense to combine your metrics data with your logs (because you want to find out why that metric is high/low and it doesn't always make sense to duplicate your data if you have a ton of it) and Elasticsearch is already a great search/analytics engine for logs. So Elasticsearch is increasingly being used as a backend for IoT systems.

The above just sort of begins to touch on it. Elasticsearch is used as the backbone for eDiscovery applications like kCura, searching Electronic Medical Records (EMRs) at places like UCLA, tracking and analyzing stock trades at Goldman Sachs, powering advertising platforms, eCommerce websites, social media analytics and NLP engines, and many other use cases.

Because of the variety of use cases, the third way to answer is just to turn it around and ask you "what do you plan on doing?"