Thank you Gehard,

Actually it happens in the middle of the execution of a pipeline (1069 rows 
have been already processed) so we are sure that the input data rows are 
present.
Over the week end I run again, same source code, same data and it worked well. 

I have now set the "Error Description" field name in the  "Error Handling"  
arrow. Let see if this error occurs again.

Have a nice day
Michel

-----Original Message-----
From: MITTERLECHNER Gerhard via users <[email protected]> 
Sent: Monday, 5 August 2024 08:57
To: [email protected]
Subject: AW: [EXTERN]REST Client sometime is not calling the service

Hi!

Probably already known, but just to go for sure:
when I started working with the REST client, non-executing REST requests were 
due to not having a row already in the stream. So we introduced dummy rows via 
Generate Rows or Data Grid in order to trigger the REST client.

Might this issue be relevant in your case?

Yours,
Gerhard



-----Ursprüngliche Nachricht-----
Von: [email protected] <[email protected]> 
Gesendet: Freitag, 2. August 2024 10:39
An: [email protected]
Betreff: [EXTERN]REST Client sometime is not calling the service

[Sie erhalten nicht häufig E-Mails von [email protected]. Weitere 
Informationen, warum dies wichtig ist, finden Sie unter 
https://aka.ms/LearnAboutSenderIdentification ]

Hello,

I'm observing sometimes (rarely, but it happens) that by calling a "REST 
Client" transform the call doesn't work(seems not to be done at all) in the 
pipeline.
It means despite the Output field "Result field name" & "HTTP status code field 
name" are set in the "REST Client" Transform you won't get values there (those 
two fields won't be populated) because the service is not called at all.

Every time I have a "REST client" transform I also have an "Error Handling" Hop 
which point to an "Abort" that stops the pipeline directly (I shall probably 
set an option "Abort threshold" to 10 instead of 0 but this is another story) 
transform.
So I know precisely the call parameters and can perform the call manually (the 
call can be done and brings the expected result == data look ok) Also this 
happened several times (rarely but it is annoying and randomly) (with the same 
set of data) I've checked the logs of the services : no error there (so service 
seems not to be called at all) and I'm quite confident with the data (and code) 
because if I run again a next time it will work (also with the information that 
I got in the log, a manual call will return the expected result).
Another point is that this "REST client" is called multiple times within the 
same pipeline and failed in the middle of the process (failed not with the 
first record for example)

MSDOS console provides the same log information as the one coming from the UI 
No memory exception (java heap ...) in the DOS console No parallel Process 
within Apache Hop No other tools working in the background (other windows 
applications) I have just observed this morning that we can also set Properties 
in the "Error Handling" Hop especially the "Error Description".

I have two questions :
- is this a known issue?
- How can I have details from Apache Hop about why the call could not be 
performed (currently the log is produced by the "Abort" transform only and 
provides information about the input data only , no indication about what was 
the error) ?
  [I will try to set Error Description field in the "Error Handling" Hop]

Thank you in advance for your help + Have an nice week-end Michel

Context
- Apache Hop Desktop 2.5.0 (I can't easily update to the latest version due to 
some known bug in the latest version)
- Services are in the cloud
- Desktop application hosted in the could within the same network as the 
services one
- Workflow started with the error level ERROR

Reply via email to