[ML][backport] Warn for model load failures if they have a status cod…

…e <500 (elastic#113410) * [ML][backport] Warn for model load failures if they have a status code <500 * On model load failure, warn if the failure status code was less tahn 500 * Update docs/changelog/113280.yaml # Conflicts: # x-pack/plugin/ml/src/main/java/org/elasticsearch/xpack/ml/inference/assignment/TrainedModelAssignmentNodeService.java * Update docs/changelog/113410.yaml * fix merge
thecoop · Sep 24, 2024 · 20d56d0 · 20d56d0
1 parent 6ee49a3
commit 20d56d0
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 1 deletion.
diff --git a/docs/changelog/113410.yaml b/docs/changelog/113410.yaml
@@ -0,0 +1,5 @@
+pr: 113410
+summary: "[ML][backport] Warn for model load failures if they have a status code <500"
+area: Machine Learning
+type: bug
+issues: []
diff --git a/...va/org/elasticsearch/xpack/ml/inference/assignment/TrainedModelAssignmentNodeService.java b/...va/org/elasticsearch/xpack/ml/inference/assignment/TrainedModelAssignmentNodeService.java
@@ -9,6 +9,7 @@
 
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
+import org.elasticsearch.ElasticsearchException;
 import org.elasticsearch.ResourceNotFoundException;
 import org.elasticsearch.action.ActionListener;
 import org.elasticsearch.action.search.SearchPhaseExecutionException;
@@ -753,7 +754,11 @@ private void updateStoredState(String deploymentId, RoutingInfoUpdate update, Ac
     }
 
     private void handleLoadFailure(TrainedModelDeploymentTask task, Exception ex) {
-        logger.error(() -> "[" + task.getDeploymentId() + "] model [" + task.getParams().getModelId() + "] failed to load", ex);
+        if (ex instanceof ElasticsearchException esEx && esEx.status().getStatus() < 500) {
+            logger.warn(() -> "[" + task.getDeploymentId() + "] model [" + task.getParams().getModelId() + "] failed to load", ex);
+        } else {
+            logger.error(() -> "[" + task.getDeploymentId() + "] model [" + task.getParams().getModelId() + "] failed to load", ex);
+        }
         if (task.isStopped()) {
             logger.debug(
                 () -> format(