# OneAgent SDK Python - IncomingWebRequestTracer - availibility to define start/response time

**URL:** <https://community.dynatrace.com/t/oneagent-sdk-python-incomingwebrequesttracer-availibility-to-define-start-response-time/47625>\
**Category:** Open Q&A\
**Tags:** dynatrace-managed, services-classic, oneagent, oneagent-sdk\
**Created:** [September 12, 2024, 1:09pm UTC](https://community.dynatrace.com/t/oneagent-sdk-python-incomingwebrequesttracer-availibility-to-define-start-response-time/47625 "2024-09-12T13:09:05Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![cscotti78](https://avatars.discourse-cdn.com/v4/letter/c/f17d59/32.png) [@cscotti78](https://community.dynatrace.com/u/cscotti78)\
**Post date:** [September 12, 2024, 1:09pm UTC](https://community.dynatrace.com/t/oneagent-sdk-python-incomingwebrequesttracer-availibility-to-define-start-response-time/47625/1 "2024-09-12T13:09:05Z")

</div>

Hi  
For some Dynatrace services, we have to recontruct the call from logs and simulate the processing time. This has limitation when the log came to late. We notice that&nbsp;IncomingWebRequestTracer keep the ingestion time for “client-side start time”. It is not that we want. The trace is not correct because it is due to recontruct service ingestion process after fetching the logs. And do not reveal the correct time.

Is it possible to specify ourself the following parameter :  
- client-side start time

- client-side response time

- server-side start time

-server-side reponse time  
- processing time

 ![cscotti78_1-1726145538451.png](https://global.discourse-cdn.com/dynatrace/original/3X/7/0/7099a68feefa9cfc8b2fa9e8c88e0ce282783653.png)

If not, I will post a request in the OneAgent SDK Python (for&nbsp;trace\_incoming\_web\_request)

---

<div class="post-metadata">

**Author:** ![Julius\_Loman](https://dub1.discourse-cdn.com/dynatrace/user_avatar/community.dynatrace.com/julius_loman/32/51_2.png) [@Julius\_Loman](https://community.dynatrace.com/u/Julius_Loman)\
**Post date:** [September 12, 2024, 7:49pm UTC](https://community.dynatrace.com/t/oneagent-sdk-python-incomingwebrequesttracer-availibility-to-define-start-response-time/47625/2 "2024-09-12T19:49:23Z")

</div>

No,it is&nbsp; not possible to specify timings. The client side timing is measured from the calling side (not sure what’s that in your case).

Did you consider using OpenTelemetry instead? It allows you to define exact start/end times.

---

<div class="post-metadata">

**Author:** ![cscotti78](https://avatars.discourse-cdn.com/v4/letter/c/f17d59/32.png) [@cscotti78](https://community.dynatrace.com/u/cscotti78)\
**Post date:** [September 12, 2024, 8:27pm UTC](https://community.dynatrace.com/t/oneagent-sdk-python-incomingwebrequesttracer-availibility-to-define-start-response-time/47625/3 "2024-09-12T20:27:55Z")

</div>

Do you think that it is possible to find a workarround with opentelemetry in order to backdate (less than 24h) a http call and process it as a service in Dynatrace ?

---

<div class="post-metadata">

**Author:** ![Julius\_Loman](https://dub1.discourse-cdn.com/dynatrace/user_avatar/community.dynatrace.com/julius_loman/32/51_2.png) [@Julius\_Loman](https://community.dynatrace.com/u/Julius_Loman)\
**Post date:** [September 12, 2024, 8:59pm UTC](https://community.dynatrace.com/t/oneagent-sdk-python-incomingwebrequesttracer-availibility-to-define-start-response-time/47625/4 "2024-09-12T20:59:56Z")

</div>

Yes, but it’s just 60 minutes in the past (see [limits](https://docs.dynatrace.com/docs/shortlink/opentelemetry-traces-ingest#ingestion-limits)), but it allows you to specify start and end times. But Python OpenTelemetry API has options for both.

I’m not familiar with your environment and requirements, but I’d consider skipping opentelemetry api/sdk completely and producing compatible json payloads instead ([example](https://github.com/open-telemetry/opentelemetry-proto/blob/main/examples/trace.json)). You can have some inspiration [here](https://github.com/agardnerIT/tracepusher).

---

<div class="post-metadata">

**Author:** ![cscotti78](https://avatars.discourse-cdn.com/v4/letter/c/f17d59/32.png) [@cscotti78](https://community.dynatrace.com/u/cscotti78)\
**Post date:** [October 14, 2024, 4:35pm UTC](https://community.dynatrace.com/t/oneagent-sdk-python-incomingwebrequesttracer-availibility-to-define-start-response-time/47625/5 "2024-10-14T16:35:01Z")

</div>

Hi,  
Thanks for your response.

I have in parallel submit a feature request on python one agent github projet.

> <https://github.com/Dynatrace/OneAgent-SDK-for-Python/issues/38>
>
> Hi, we try to reconstruct services base on our logs for saas application. It wor…ks fine with the SDK. The problem is that the one agent will put itself the start and end timestamps. It could be wrong if the logs are parse and process late especially if they are some spicks of logs (5 or 20 minutes delay, even more).
> 
> As in open telemetry, could you provide a possibility to let us define manually the correct start\_time and end\_time (timestamps) on :
> \- trace\_incoming\_web\_request, 
> \- trace\_outcoming\_web\_request, 
> \- trace\_custom\_service.
> 
> We find a by pass with open telemetry. But with this solution, we lost the service flow visualisation.
> 
> If possible, we would like to not implement hybride ingestion solutions (One Agent + Open Telemetry) in order to have an expected distributed traces with all the features that dynatrace could provide (based on One Agent ingestion).

Regards,  
Carlo
