Downgrade Splunk HEC error to warning (#732)

* Downgrade Splunk HEC error to warning * Update CHANGELOG
stripe · Jun 27, 2019 · 9d9345e · 9d9345e
1 parent 3c3bc47
commit 9d9345e
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 1 deletion.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -13,6 +13,7 @@
 ## Updated
 
 * Updated the vendored version of DataDog/datadog-go which adds support for sending metrics to Unix Domain socket. Thanks, [prudhvi](https://github.com/prudhvi)!
+* Splunk sink: Downgraded Splunk HEC errors to be logged at warning level, rather than error level. Added a note to clarify that Splunk cluster restarts can cause temporary errors, which are not necessarily problematic. Thanks, [aditya](https://github.com/chimeracoder)!
 * Updated the vendored version of github.com/gogo/protobuf which fixes Gopkg.toml conflicts for users of veneur. Thanks, [dtbartle](http://github.com/dtbartle)!
 
 ## Bugfixes

diff --git a/sinks/splunk/splunk.go b/sinks/splunk/splunk.go
@@ -401,7 +401,7 @@ func (sss *splunkSpanSink) makeHTTPRequest(req *http.Request, cancel func()) {
 			"hec_status_code":   parsed.Code,
 			"hec_response_text": parsed.Text,
 			"event_number":      parsed.InvalidEventNumber,
-		}).Error("Error response from Splunk HEC")
+		}).Warn("Error response from Splunk HEC. (Splunk restarts may cause transient errors).")
 	}
 	samples.Add(ssf.Count(failureMetric, 1, map[string]string{
 		"cause":       cause,