APM Server specification- Recommendations

Elastic Observability APM
1

Can anybody help me in getting below information: -

  1. APM Server Specification
  2. APM Recommendations
2

These are very broad questions. Can you please explain what your goal is, describe what you've tried and which docs you've read, and explain what information you're still missing? That will help us point you in the right direction.

If you didn't get as far as finding and reading the APM Server docs, you can find them here: APM Server Reference [7.12] | Elastic

3

Hey! Thanks for your reply :slight_smile:
I am actually interested in knowing the recommendations for APM server such as CPU, memory, disk usage etc. which I am not able to find in any document/blog.

4

It depends on your environment. How many instrumented services? What's the request rate? Are there spikes? etc. To get you started, take a look at these pages:

If you can provide details of your environment then someone may be able to assist further. Either way, I'd recommend experimenting.

1 Like
5

Hi @axw,
I went thorugh the documentation for APM sizing and performance guide and it was helpful.
But I want some more information on the CPU cores and other specifications irrespective of what environment we are working in.
Could you please let me know if there is any blog which I can refer to get the above information?

6

One more question regarding APM!
When our elasticsearch cluster is down, APM server buffers data temporarily. In APM, is there any possibility to enable persistance storage so that we don't loose any data. Similar to logsatsh.

7

But I want some more information on the CPU cores and other specifications irrespective of what environment we are working in.
Could you please let me know if there is any blog which I can refer to get the above information?

I'm not aware of any documentation or blogs that go to that level of detail.

When our elasticsearch cluster is down, APM server buffers data temporarily. In APM, is there any possibility to enable persistance storage so that we don't loose any data. Similar to logsatsh.

Currently there is no supported feature for this. I have opened an issue which you can track: Support spooling to disk · Issue #5441 · elastic/apm-server · GitHub

1 Like
8

Hi @Nisha2297

You can see what HW profiles we use in Elastic Cloud here

If you are on 7.13 +

I would suggest a minimum of 1GB RAM because APM server is also Fleet server.

If you are going to do higher volumes scale up to

2,4 GB : 1+ vCPU
8 GB and 2+ vCPU

1 Like
9

This topic was automatically closed 20 days after the last reply. New replies are no longer allowed.