Issues with Signoz 10.2: Grouping, Missing Traces, and Capturing Requests

TLDR Anil reported issues with Signoz 10.2, including grouping, missing traces, and capturing requests. Srikanth suggested upgrading to a newer version. The issue was resolved after upgrading the node version to 18 and additional clarifications were provided.

Photo of Anil
Anil
Thu, 13 Apr 2023 18:12:55 UTC

Hi Team, I am using signoz 10.2 integrated into an API server and i am able to monitor the API server. But I observe few things below 1. multiple spans are grouping together (of different requests) 2. i can see lot of missing traces recently. 3. when I tried to submit the same request API for a couple times say like 18-20 times, and tracing is showing for 8-9 times only. may i know why couple of requests are not being captured. Is there any way to solve this issue ? I kindly request you to look into it. Even i am willing connect directly to give you exact problem statement

Photo of Anil
Anil
Fri, 14 Apr 2023 11:20:49 UTC

Hi Team any update on this ?

Photo of Srikanth
Srikanth
Fri, 14 Apr 2023 11:59:22 UTC

> 1. multiple spans are grouping together (of different requests) What do you mean by this “grouping”? > 1. when I tried to submit the same request API for a couple times say like 18-20 times, and tracing is showing for 8-9 times only. may i know why couple of requests are not being captured. Is there any way to solve this issue ? This seems like an issue from SDK/instrumentation. Which language & framework are you using? What does SDK setup look like? P.S: 10.2 is pretty old and we have shipped several fixes and enhancement so you may want to upgrade and test on new version.

Photo of Anil
Anil
Mon, 17 Apr 2023 14:07:27 UTC

Srikanth, Thanks alot for your reply. we are using NodeJS instrumentation and Tomcat Instrumentation

Photo of Anil
Anil
Mon, 17 Apr 2023 14:16:52 UTC

Unfortunately, we cant go for upgrade of signoz as it lies with Client.

Photo of Anil
Anil
Thu, 20 Apr 2023 12:06:42 UTC

Srikanth , is there any further update related to this ?

Photo of Anil
Anil
Fri, 21 Apr 2023 07:56:07 UTC

Hi Team any update on this ?

Photo of Srikanth
Srikanth
Fri, 21 Apr 2023 08:13:14 UTC

You are using a very old version. We have made many fixes, including the missing spans related. Could you share a reproducible example? What version of opentelemetry are you using?

Photo of Anil
Anil
Fri, 21 Apr 2023 08:26:49 UTC

"@opentelemetry/api": "^1.1.0", "@opentelemetry/auto-instrumentations-node": "^0.28.0", "@opentelemetry/exporter-otlp-grpc": "^0.26.0", "@opentelemetry/sdk-node": "^0.28.0", and java agent version is 1.16.0

Photo of Srikanth
Srikanth
Fri, 21 Apr 2023 08:31:49 UTC

You are like 10 releases behind, which makes it really hard to help. Please update the SigNoz and OpenTelemetry versions and let us know if the issue is still not fixed.

Photo of Anil
Anil
Fri, 21 Apr 2023 10:32:52 UTC

after upgrading do i need to make changes to my node js instrumentation with latest npm pkgs?

Photo of Anil
Anil
Fri, 21 Apr 2023 11:13:43 UTC

Srikanth could you please let me know on the above query ?

Photo of Srikanth
Srikanth
Fri, 21 Apr 2023 15:20:24 UTC

> after upgrading do i need to make changes to my node js instrumentation with latest npm pkgs? Yes, you may not have to do it for Java apps but you will need to update the node packages.

Photo of Anil
Anil
Mon, 24 Apr 2023 07:15:09 UTC

thanks

Photo of Anil
Anil
Wed, 10 May 2023 09:58:48 UTC

Hi Srikanth, we have deployed latest versoin 18.1, but signoz trace is still showing missing

Photo of Anil
Anil
Wed, 10 May 2023 10:00:22 UTC

Photo of Srikanth
Srikanth
Wed, 10 May 2023 10:11:36 UTC

Then there is an issue with your instrumentation. Please share a simple reproducible example.

Photo of Anil
Anil
Wed, 10 May 2023 10:13:14 UTC

just i called a rest api from post man .

Photo of Anil
Anil
Wed, 10 May 2023 10:13:57 UTC

`traceExporter = new OTLPTraceExporter();` sdk = new opentelemetry.NodeSDK({ traceExporter, instrumentations: [getNodeAutoInstrumentations(), new MongoDBInstrumentation({ enhancedDatabaseReporting: true, }),], }); // initialize the SDK and register with the OpenTelemetry API // this enables the API to record telemetry sdk.start()

Photo of Anil
Anil
Wed, 10 May 2023 10:14:39 UTC

if possible for you we can huddle

Photo of Srikanth
Srikanth
Wed, 10 May 2023 10:20:30 UTC

If you share a simple reproducible example, it can be debugged further to see where the issue is; otherwise any idea of getting on to call doesn’t make sense. How can call help if we don’t know where it is coming from?

Photo of Anil
Anil
Wed, 10 May 2023 10:21:53 UTC

Ok could you please more specific on the sample ?

Photo of Srikanth
Srikanth
Wed, 10 May 2023 10:22:43 UTC

It can be any simple node app that replicates the issue you mentioned.

Photo of Anil
Anil
Wed, 10 May 2023 10:50:36 UTC

ok

Photo of Anil
Anil
Wed, 10 May 2023 11:53:22 UTC

if i can give you github link will that be fine ?

Photo of Srikanth
Srikanth
Wed, 10 May 2023 11:58:06 UTC

Yes

Photo of Anil
Anil
Mon, 15 May 2023 08:26:25 UTC

Srikanth, we have resolved the issue by upgrading the node version to 18. Could you please help us with document for 18.1 as the site is providing documentation for latest version?

Photo of Anil
Anil
Mon, 15 May 2023 12:34:44 UTC

Also Srikanth, please see the below picture. in the picture you will see 450 ms of POST request, but below spans shows that it took 56 ms to complete the request . I would like to know the remaining time 380ms details. Could please help me to understand this

Photo of Anil
Anil
Tue, 16 May 2023 06:00:52 UTC

Hi Team , Any update on this ?

Photo of Anil
Anil
Tue, 16 May 2023 06:55:50 UTC

Hi <@4K143e> /Srikanth, could you please help me out ?

Photo of Srikanth
Srikanth
Tue, 16 May 2023 15:33:30 UTC

> Could you please help us with document for 18.1 as the site is providing documentation for latest version I am not sure which document you are referring to. > in the picture you will see 450 ms of POST request, but below spans shows that it took 56 ms to complete the request . I would like to know the remaining time 380ms details. Could please help me to understand this The sum of the child spans doesn’t always add up to the parent span duration. It doesn’t mean it took 56ms to complete the request. It means the operations till /engine-rest/… tok 56 millis, there could be another work that is not traced after which is responsible for the rest of 380ms.