Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

possibility to create a param with SpanKindServer #15

Open
jmmc-tools opened this issue Aug 21, 2023 · 3 comments
Open

possibility to create a param with SpanKindServer #15

jmmc-tools opened this issue Aug 21, 2023 · 3 comments

Comments

@jmmc-tools
Copy link

span.SetKind(ptrace.SpanKindClient)

Here only is possible to create ParameterizedGenerator with SpanKindClient.. any update to create spans kind server?





@zalegrala
Copy link
Contributor

I'm sure that could be adjusted, but can you say more about the use case? Does this impact your load test in a way?

@jmmc-tools
Copy link
Author

Yes it's because our traces in prod environment, usually have a root span kind Server and multiple nested spans of kind client. It's only to try reproduce the real traces as same as possible. But the actual behavior is good for default settings and in certains load testing could be interesting to create a param o config to enable a combination of span kind server and clients.

Another suggestions is related with number of spans in each trace. It's to find the exact limit of backend (Grafana Tempo). We have discovered high memory consumption when 1 trace have hundreds of spans, and we have received 200 or 300 traces/sec with this high level of spans per trace, Tempo use about 128 GB of Ram (monolitic installation in k8s). So specify the spans per trace, or spans per second could be a good feature in this k6 plugin to determinate that hard limit.

Thanks!





@zalegrala
Copy link
Contributor

I think the default param.js should be modified to reduce the number of attributes. Tweak as necessary for your environment. This is likely where most of the memory is going. I've been able to achieve 1k spans per second on about 12G of memory with some tuning an small 3 ingested environment. We probably need some changes here to improve the configuration, but not sure that someone will pick this up soon. If you'd like to send PRs, we'd be happy to review.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants