Hi All. Great product. Keep up the great work please and continue to "push" into use cases not just for focus of log analysis...but please push the envelope for those like my project that have dumped SQL databases for noSQL ES.
My scenario.... we are ingesting a 30-table midddleware product/tools data into ElasticSearch so we can give live visualization of the active data that this product/tool is not doing well in terms of visualizing trends on metrics and viewing metdata for 'things' that are chronic.
I'm sorry but i have never seen any strong seminar recording or write up on the methodology/best-practice in how one would take a 30+ SQL table source that speaks of around 5 critical major entities and their relationships and how best to build index structures for maximizing what can be done from one dashboard and having the related data alongside each major entity. Even if the solution is ...."Demormalize and Create the widest, flattest document of all fields of all related entities even if massively redundancy will exist"..... i want to hear that that is the best practice.
I hear about this 'nested' stuff but i don't want to end up with something where only a few people can possibly figure out how to navigate such data in Discover.
Can someone refer me to best practices for how best to land SQL data that is highly normalized into ES ? I'm just looking for the 'token' article, seminar or writeup.
Just feels like over the last 2 years people write on the sides of this but no one is really helping from Elastic to give a strong writeup for those like my team that are using Elastic for far more than time series logs and metrics but bringing in other structured data that is metadata about entities. VMs, their hypervisor managers, their location data, their relationship data between data stores and storage volumes.... not just times series..... metrics and metdata about taxonomies of the "things" we are getting metdata for.
Love the product, hopeful someone can help me.