New to Elastic Search - General question searching through a tagged text file

I have a general question about Elasticsearch / Kibana and if it is the right tool for the job I am trying to do. Apologies in advance if this is the wrong place for this or if the post is too long.

I have a file that contains network device outputs such as version, network connections, power usages, etc, etc. When I generate the file, my script places the name of the device and the command on every line of output. It looks something like this:

DeviceA, show version, v1.0
DeviceA, show connections, connected to deviceX
DeviceA, show connections, connected to deviceY
DeviceB, show version, v2.0
DeviceB, show connections, connected to deviceAlpha
DeviceB, show connections, connected to deviceBeta

What I end up with is a searchable file of state information for a large infrastructure. Currently I search this file you using, grep, sed, awk, uniq, wc (word count) etc. So for example, to check what devices have the most connections, I would do the following: grep "connected to" | print the first column only using awk | then count all the uniq entries. In this case I would end up with DeviceA - 2, and DeviceB -2. Or I could just grep "show version" and see the version of every device.

With this tagged file, unix search, and filters, I can search my entire infrastructure for an almost limitless combination of variables. My only problem is that nobody will use it because it is command line and not remotely resembling a product.

It would be great if anyone has an opinion if Elasticsearch and Kibana would work for this. Also it would be great to know if Logstash is required as these are static files. If this works, Elastic could make our entire infrastructure like a searchable pivot table. Servers, Storage, Databases could all be easily added.

Finally, advice on how to get the required stack up and running as quickly as possible would be great. I could have cloud provider as a proof of concept but would eventually probably host on premise.

Thanks!

Welcome :slight_smile:

It sounds like the Elastic Stack would be a good fit for this. It looks like Filebeat and an Ingest pipeline would do what you want pretty easily. Logstash is good for doing more advanced processing of data, and you can add that to the mix later on if you want.

If you want to get started quickly, then our Elasticsearch Service will offer hosted Elasticsearch and Kibana, with a 14 day free trial to start (and if you would like an extended trial, drop me a PM and I should be able to send you a link). Otherwise we offer docker containers of Elasticsearch and Kibana which are a good way of running up on prem quickly.

Thanks Mark. I think I will try the service. One question. Would I be using Kibana to do the searching using a web interface?

Because the entire reason for doing this would be to disseminate "searching the infrastructure" to a wider audience. I can already search the text file in an unlimited combination of ways using unix commands but everyone else it deathly afraid of unix.

Yes, Kibana is the interface for that sort of thing.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.