There is a question in front of me, Why we go to elasticseach+Kibana for visualization? There are some other techniques available for the same like Grafana+InfluxDB. Suppose my answer is "ES+Kibana is better way than others", How can I justify my answer?
Have a look at this awesome thread:
I have exactly the same situation as you are describing.
Grafana+Influxdb or "ES+Kibana"
I don't think there is an answer for what's a better option, it depends on your environment, in my case I ended up with ES+Kibana+Grafana for the following reasons:
Influxdb is not stable, I couldn't get it to ingest data into a cluster for a few hours without some sort of problem. With ES I only had to tweak to not overload the system.
ES can plug into Grafana AND Kibana, so I had the best of both worlds. As Grafana does not have as many plotting options as Kibana.
I have no problems with storage capacity for the amount of data I need to handle nor ingestion rates that ES can't cope with (in the case that influxdb could ingest more metrics per second).
ES is much more stable and mature, it may have not been designed as a timeseries DB, but it does the job well as far as I can tell and it's being used in many places for this.
You also have timelion in case you don't want to use grafana and keep everything under kibana.
hope that helps,
Thank you very much Isaac. This is what I expect from this thread .
Also check german23 link, it has a more in depth explanation of it. Even though it's a bit old and grafana wasn't available for ES, the storage point they make are valid
I'm using ELK, InfluxDB and Grafana and there're just my two cents: Grafana as single centrilized dashboard in your company seems better option because of DataSource feature.
With Grafana dashboard you can draw graphs from multiple DataSource: ElasticSearch, InfuxDB, Prometheus, etc. You can use different ElasticSearch clusters as DataSource on one graph whereas KIbana works with only one ElasticSearch cluster and cannot visualize data from another TSDB.
So Grafana is much more universal as vusualizing tool.