Merge pull request #162 from posit-dev/fix-alloy-blackbox-instances

fix: alloy instance duplication bug

Author: amdove

python-pulumi/src/ptd/pulumi_resources/grafana_alloy.py

@@ -493,12 +493,27 @@ def _define_config_map(
 
                 prometheus.scrape "blackbox" {{
                     targets    = prometheus.exporter.blackbox.front_door.targets
-                    forward_to = [prometheus.relabel.default.receiver]
+                    forward_to = [prometheus.relabel.blackbox.receiver]
                     clustering {{
                         enabled = true
                     }}
                 }}
 
+                // Normalize instance label for blackbox metrics to deduplicate across Alloy pods.
+                // Each pod runs its own blackbox exporter, producing metrics with different
+                // instance labels (node hostnames). Replacing instance with a static value
+                // ensures Mimir deduplicates them, preventing duplicate alerts.
+                // See: https://github.com/grafana/alloy/issues/1009
+                prometheus.relabel "blackbox" {{
+                    forward_to = [prometheus.relabel.default.receiver]
+
+                    rule {{
+                        action       = "replace"
+                        target_label = "instance"
+                        replacement  = "blackbox"
+                    }}
+                }}
+
                 prometheus.scrape "pods" {{
                     targets    = discovery.relabel.pod_metrics.output
                     forward_to = [prometheus.relabel.default.receiver]