Metricbeat Haproxy dashboards

Elastic Stack Beats
1

Hello Team,

I have configured Metricbeat Haproxy Module and i am getting the data over kibana dashboard under Metricbeat Haproxy Dashboards.

I am able to see 6 dashboards for only Metricbeat Haproxy which are listed below:

  1. [Metricbeat HAProxy] Overview
  2. [Metricbeat HAProxy] HTTP server
  3. [Metricbeat HAProxy] HTTP backend
  4. [Metricbeat HAProxy] Frontend
    5 [Metricbeat HAProxy] Backend
    6 [Metricbeat HAProxy] HTTP frontend

Some values are showing in multiple dashboards and i have tried to understand the graph on all dashboards. But i am not getting some values and graph.

Can any body provide me any link which described the Values of all graph? It will be quite easy for me to understand all the graph and values.

Thanks.

2

Hi @Tek_Chand,

HAproxy can be used for many things and allows a variety of configurations, this is why we decided to provide a variety of dashboards that can be used for different configurations, or whose visualizations can be reused for custom dashboards for specific use cases. Some graphs won't show any data if you don't use some haproxy features, for example you won't see anything on HTTP visualizations if HAproxy is used as TCP proxy only.

Is there something in particular that you expect to appear in these dashboard and you cannot find?

There is no additional documentation at the moment, you can take a look to the original PR (#6934). Also feel free to open a new issue to request more documentation about these dashboard.

3

@jsoriano, Thank you for your response.

I need to make deep dive in Haproxy to understand all dashboards. Currently i have only basic idea of Haproxy thats why its quite difficult for me to understand all the dashboards.

If i need to customise any dashboard i will get back to you.

Thanks.

4

@jsoriano,

I am checking [Metricbeat HAProxy] Backend dashboard and there is graph for Average connection time and its self explanatory its shows it for last 1024 requests. But there are 3 values in percentile. Can you please help me to understand what are these values. Please refer the below screenshot:

Selection_009
Selection_009.png838×228 26.6 KB

When we check the [Metricbeat HAProxy] Frontend for last 15 minutes interval or more graph start with very high values and we can only understand when we read the values from right hand side. Please refer the below screenshot:

Selection_010
Selection_010.png1657×449 52.3 KB

Thanks.

5

Now thinking on this these percentiles don't make much sense to me. A percentile indicates the value below which a percentage (99, 90, 50...) of metrics are included. But here this is being calculated not from the raw data but from the averages, what I don't think is correct. I will review this. I have opened an issue Haproxy visualization calculating percentiles from average · Issue #10052 · elastic/beats · GitHub

Data shown here is calculated as the derivative of a monotonically increasing value. It can show peaks if there is no previous data, or if the counter has been reset. Does it happen only on specific intervals, or always?

6

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