Elastic Uptime - Best Practices

Hello Dear Community,

I'm planning to set up Elastic Uptime in three different locations to monitor the response times of services.

The goal is to track how the response times of services in our data centers change over time.

We aim to identify potential network or application-related delays in real-time and receive alerts when such delays occur.

I would appreciate guidance on best practices for implementing this monitoring setup

Thanks in advance!

Hi @serkol,

Welcome to the community and thank you for using Elastic Uptime!

We've been discussing this within the team and there are a few questions/statements:

  • What version of Kibana are you using?
  • Are you using elastic cloud or using self-managed instead?

We believe that the best approach if you are using the cloud, is to use 3 public locations and to use 3 private locations if you are not.

Cheers,
Alberto

Hello Alberto,

Thanks for reaching me out!
We haven't set the architecture up yet.

We are planning to use the latest version of Kibana.
We are planning to use self-managed.

We'll have 3 private locations.

Can you kindly tell me more about the best-practices? :slight_smile:

Thanks in advance!

Hi @serkol,

It's a broad topic, so please let us know if there's anything specific you'd like to know, in order to give you better feedback.

If are interested in the "configuration as code" approach, I'd recommend you look into our project monitor initiative. With these, you'll be able to manage your monitors as if it were a deploy-able artifact instead of manually setting them up through the UI. Both options are supported, but project monitors are a better DX experience.

Since you'll be using self-hosted, please also consider our scalability guidelines for private locations. Scaling will be directly correlated to the amount of monitors you're expecting to run so you might to fine-tune resource allocation for these.

If data retention is a point of concern, I'd recommend reading through this doc section and follow this tutorial if you need to customize any of them. Other than browser monitors' indeces, data volume is fairly minimal on average across tcp, icmp and http monitor types.

As additional topics, we have this guide on managing secrets/credential inside synthetics and this statement on how data is secured on clusters.

Hope that helps!

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