Add status reports for HDFS and HBase

docs/latest/modules/en/attachments/suse-observability_logs_collector.sh

@@ -1,6 +1,6 @@
 #!/bin/bash
 
-ELASTICSEARCH_LOGS=0
+ELASTICSEARCH_LOGS=false
 ELASTICSEARCH_RANGE="7d"
 while getopts "her:" option; do
   case $option in
@@ -23,7 +23,7 @@ options:
 EOF
       exit 0;;
      e) # Collect elasticsearch logs
-      ELASTICSEARCH_LOGS=1;;
+      ELASTICSEARCH_LOGS=true;;
      r) # Time range for elasticsearch logs
       ELASTICSEARCH_RANGE=$OPTARG;;
     \?) # Invalid option
@@ -49,12 +49,12 @@ for cmd in ${COMMANDS[@]}; do
 done
 
 # skip helm release analysis when not all its dependencies are present
-HELM_RELEASES=1
+HELM_RELEASES=true
 for cmd in base64 gzip jq
 do
   if ! command -v $cmd &>/dev/null; then
      echo "$cmd is not installed. Skipping analysis of helm releases."
-     HELM_RELEASES=0
+     HELM_RELEASES=false
   fi
 done
 
@@ -130,11 +130,25 @@ collect_pod_disk_usage() {
 collect_helm_releases() {
     techo "Collecting helm releases..."
     mkdir -p "$OUTPUT_DIR/releases"
+
+    # Restrict keys extracted from Helm values to only this include-list to avoid including any
+    included_keys='["resources", "affinity", "nodeSelector", "tolerations"]'
+
+    # 1. --argjson keys "$included_keys": Passes the shell variable as a JSON array $keys.
+    # 2. . as $input: Saves the entire original JSON into a variable $input.
+    # 3. [ paths | ... ]: Gathers all paths from the JSON.
+    # 4. select(.[-1] as $last | $keys | index($last)): Selects only paths where
+    #    the last element (.[-1]) is found inside the $keys array.
+    # 5. reduce .[] as $p (null; ...): Starts with an empty (null) document
+    #    and iterates over every path ($p) that was selected.
+    # 6. setpath($p; $input | getpath($p)): For each path, it sets that path
+    #    in the *new* document, pulling the *value* from the original $input.
+
     RELEASES=$(kubectl -n "$NAMESPACE" get secrets -l owner=helm -o jsonpath="{.items[*].metadata.name}")
     for release in $RELEASES; do
         kubectl -n "$NAMESPACE" get secret "$release" -o jsonpath='{.data.release}' | \
           base64 --decode | base64 --decode | gzip -d | \
-          jq '{ info: .info, metadata: .chart.metadata, config: .config }' > "$OUTPUT_DIR/releases/$release"
+          jq --argjson keys "$included_keys" '{ info: .info, metadata: .chart.metadata, config: ( .config as $input | [ .config | paths | select(.[-1] as $last | $keys | index($last)) ] | reduce .[] as $p (null; setpath($p; $input | getpath($p)))) }' > "$OUTPUT_DIR/releases/$release"
     done
 }
 
@@ -247,6 +261,35 @@ EOF
     kill $CHILD
 }
 
+collect_hdfs_report() {
+  POD=$(kubectl -n "$NAMESPACE" get pod -l app.kubernetes.io/component=hdfs-nn -o jsonpath='{.items[0].metadata.name}' 2>/dev/null) || true
+  if [ -n "$POD" ]; then
+    techo "Collecting HDFS report..."
+    mkdir -p "$OUTPUT_DIR/reports"
+    kubectl exec -n "$NAMESPACE" "$POD" -c namenode -- bash -c "unset HADOOP_OPTS; hdfs dfsadmin -report" > "$OUTPUT_DIR/reports/hdfs.log"
+  fi
+}
+
+collect_hbase_report() {
+  POD=$(kubectl -n "$NAMESPACE" get pod -l app.kubernetes.io/component=hbase-master -o jsonpath='{.items[0].metadata.name}' 2>/dev/null) || true
+  if [ -n "$POD" ]; then
+    # Running in HA Mode
+    techo "Collecting HBase report..."
+    mkdir -p "$OUTPUT_DIR/reports"
+    kubectl exec -n "$NAMESPACE" "$POD" -c master -- bash -c 'hbase hbck -details 2>&1' > "$OUTPUT_DIR/reports/hbase.log"
+  else
+    POD=$(kubectl -n "$NAMESPACE" get pod -l app.kubernetes.io/component=stackgraph -o jsonpath='{.items[0].metadata.name}' 2>/dev/null) || true
+    if [ -n "$POD" ]; then
+      # Running in non-HA mode
+      techo "Collecting HBase report..."
+      mkdir -p "$OUTPUT_DIR/reports"
+      kubectl exec -n "$NAMESPACE" "$POD" -c stackgraph -- bash -c 'hbase hbck -details 2>&1' > "$OUTPUT_DIR/reports/hbase.log"
+    else 
+      techo "Could not find HBase or StackGraph pod to generate HBase report."
+    fi
+  fi
+}
+
 archive_and_cleanup() {
     echo "Creating archive $ARCHIVE_FILE..."
     tar -czf "$ARCHIVE_FILE" "$OUTPUT_DIR"
@@ -303,11 +346,13 @@ kubectl -n "$NAMESPACE" get events --sort-by='.metadata.creationTimestamp' > "$O
 # Run the pod logs collection function
 collect_pod_logs
 collect_pod_disk_usage
+collect_hdfs_report
+collect_hbase_report
 collect_yaml_configs
-if [ $HELM_RELEASES ]; then
+if $HELM_RELEASES; then
   collect_helm_releases
 fi
-if [ $ELASTICSEARCH_LOGS ]; then
+if $ELASTICSEARCH_LOGS; then
   collect_pod_logs_from_elasticsearch
 fi