Different reaction on API call with the same version of Graylog

Graylog Central (peer support)
1

Hello Together,

we have two Graylog servers. One test environment and one production environment.
Now I made an API request on the test environment like this.

https://mygraylog.test.com/api/search/universal/relative?query=*&range=18600&filter=streams:62fcfbf6c385d02010b32ded

There I get as desired a JSON with the appropriate messages back.
As soon as I do this on our production environment with the appropriate adjustment I get the message

must not be empty (path = RelativeSearchResource.searchRelativeChunked.arg6, invalidValue = )

Interestingly, if I build an API call in the API browser of the production environment and call it in the API browser, it works.

If I then paste the generated URL into the browser, the above error message appears.

The graylog version on both hosts is the same Version of Graylog

  • Package Version: 5.0.8

There are only the following differences between the two environments in the ENVs
In production the following additional parameters are

GRAYLOG_OUTPUT_BATCH_SIZE
GRAYLOG_SERVER_JAVA_OPTS
GRAYLOG_MESSAGE_JOURNAL_MAX_SIZE

The graylog.conf is identical

I checked if the index ids are correct, if the streamids are the right ones. And even if they were wrong, the API request would have to tell me that I am entering a wrong value.

Unfortunately I have no more ideas

2

These are 2 endpoints that use the same route and are mapped based on the returned data type (JSON vs CSV). The selection is done according to the Accept HTTP header.
Some clients send */* as an Accept header, in which case the resolution is not well-defined. Explicitly setting the header works to disambiguate.

3

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.