Greetings, I’d like to ask about the

Database Calls

tab having no data despite there are traces that are related to PostgreSQL queries (see 2nd screenshot) Is there any extra steps or documentations that I missed that I needed to take to gain visibility on these database metrics? I am currently using a java auto instrumentation from OpenTelemetry 1.13.1 w/ the startup settings as follows

-Dotel.metrics.exporter=none -Dotel.exporter.otlp.endpoint=<collector endpoint of signoz>:4317 -Dotel.resource.attributes=service.name=servicenamesomething -javaagent:/path/to/otel_agent.jar

Currently using Signoz installed from a helm chart with the collector version

0.43.0-0.1

Thanks in advance!

Do you see

db.system

in the trace attributes for the db calls?

Yes, there is

db.system

with the value

postgresql

Does you time range for db calls cover the db trace timestamps range?

Yes, I have also tried to set it until 1 week ago, still no results unfortunately

Interesting

Yes, tracing works normal to me Okay then, thanks, let me know anytime if you need more information!

@Steven Sim Can you help with some troubleshooting here. Can you access the clickhose and share the result of this query /

Yes, I’d be glad if I can help This are the results, I found no result w/ the

WHERE

statement

Yeah, it's unrelated

also empty unfortunately

Ok there is no data at all? This is getting mysterious.

yes, no data at all, there’s data if I remove the

WHERE

statement though

What data do you see? Would you be able to share the some metric names?

Here’s if I were to remove the

WHERE

statement second if I were to use

*

, all of the

__name____

starts with

system

Thanks, there are no metrics generated from the trace data which are essential for the

Database Calls

tab. This is probably a bug

Ok thanks!, let me know if you have any updates on this topic

@Steven Sim Can you share the environment details to reproduce this? That would be helpful.

I’m using Signoz installed on a kubernetes cluster w/ the current latest helm chart Java version 8 on AWS EC2, postgres 13.4 on AWS RDS, Opentelemetry auto instrumentation jar v1.13.1

Alright, that helps. Will look into this.

Hi @Srikanth Chekuri , sorry to bother you again but Is there any updates on this topic?, and if not yet is there a Github issue already that I can track for the upcoming time?

@Steven Sim Spent some time debugging, this error shows up inconsistently. Do you not see data for just after one request or no data at all? . When I was debugging It was observed this is an issue if I just make one to two requests but If I have an app running with near real work load I was able to see them although some of them were getting dropped as mentioned above.

In my case there was no data at all, I installed it on 3 servers running the same application and the workload was quite medium in my opinion Maybe is there a built-in sampling for these DB traces?

No there is no sampling as of now.

I’m quite sure at least 95% of the workload has something to do with postgres

Will you be able to share the logs for your collector instance?

Sure, I’ll get back to you soon