-
Bug
-
Resolution: Unresolved
-
Minor
-
None
-
Jenkins Core 2.190.1
Logstash Plugin 2.3.1
I came across an error like this:
[logstash-plugin]: Failed to send log data: <URL>
[logstash-plugin]: No Further logs will be sent to <URL>
java.io.IOException: HTTP error code: 400
RESPONSE: HttpResponseProxy{HTTP/1.1 400 Bad Request [content-type: application/json; charset=UTF-8] org.apache.http.client.entity.DecompressingEntity@3a22308c}
{"error":{"root_cause":[{"type":"illegal_argument_exception","reason":"mapper [data.buildVariables.Foo] of different type, current_type [text], merged_type [ObjectMapper]"}],"type":"illegal_argument_exception","reason":"mapper [data.buildVariables.Foo] of different type, current_type [text], merged_type [ObjectMapper]"},"status":400}
at jenkins.plugins.logstash.persistence.ElasticSearchDao.push(ElasticSearchDao.java:190)
at jenkins.plugins.logstash.LogstashWriter.write(LogstashWriter.java:172)
at jenkins.plugins.logstash.LogstashWriter.writeBuildLog(LogstashWriter.java:137)
at jenkins.plugins.logstash.pipeline.LogstashSendStep$Execution.run(LogstashSendStep.java:77)
at jenkins.plugins.logstash.pipeline.LogstashSendStep$Execution.run(LogstashSendStep.java:54)
Doing some investigation I found that one of our plugins was adding an environment variable with "." in the name like "foo.bar". When I removed those environment variables it fixed the issue. I believe the logstash plugin could and probably should handle this case in a more graceful way - even if just erroring and saying the variable that can't be sent to Elastic. Or more advanced might be replacing the "." with something else so it just works.