WIK-2147: Group log archives by year subdirectory

README.md

@@ -68,6 +68,7 @@ The image ships with `archive-cron-logs`, a log archiving script designed for th
   - Additional archives for adjacent periods are created so archived ranges do not intersect the active period.
 - **Recent logs preservation**: The script skips the newest `target_count` log files to keep the latest logs easy to read.
 - **Naming**: When all files in a group belong to the same task, the task name is appended to the archive name as shown above.
+- **Directory layout**: Archives are grouped by year in subdirectories (e.g. `2025-01-01_taskname.tar.gz` is stored as `2025/01-01_taskname.tar.gz`). The script also discovers existing archives in both the flat directory and year subdirs.
 
 You can schedule archiving via container labels, for example in `docker-compose`:
 

archive-cron-logs-functions.sh

@@ -19,14 +19,47 @@ find_newest_archive() {
         return 1
     fi
 
-    # Build timestamp:file pairs and sort newest->oldest, then take the first entry (newest)
+    # Build timestamp:file pairs and sort newest->oldest, then take the first entry (newest).
+    # Look in both flat dir ($folder/*.tar.gz) and year subdirs ($folder/*/*.tar.gz).
     local newest_archive=""
     shopt -s nullglob
     while IFS= read -r pair; do
         # Take the first entry (newest after reverse sort)
         newest_archive=$(echo "$pair" | cut -d: -f2-)
         break
     done < <(
+        for archive_file in "$folder"/*/*.tar.gz; do
+            [ -f "$archive_file" ] || continue
+            local year_subdir
+            year_subdir=$(basename "$(dirname "$archive_file")")
+            local basename_archive
+            basename_archive=$(basename "$archive_file")
+            local name_without_ext="${basename_archive%.tar.gz}"
+            local period_key="${year_subdir}-${name_without_ext}"
+            local timestamp=""
+            if [[ "$period_key" =~ ^([0-9]{4})(_.*)?$ ]]; then
+                local year="${BASH_REMATCH[1]}"
+                timestamp="${year}0101000000"
+            elif [[ "$period_key" =~ ^([0-9]{4})-([0-9]{2})(_.*)?$ ]]; then
+                local year="${BASH_REMATCH[1]}"; local month="${BASH_REMATCH[2]}"
+                timestamp="${year}${month}01000000"
+            elif [[ "$period_key" =~ ^([0-9]{4})-([0-9]{2})-([0-9]{2})(_.*)?$ ]]; then
+                local year="${BASH_REMATCH[1]}"; local month="${BASH_REMATCH[2]}"; local day="${BASH_REMATCH[3]}"
+                timestamp="${year}${month}${day}000000"
+            elif [[ "$period_key" =~ ^([0-9]{4})-([0-9]{2})-([0-9]{2})_([0-9]{2})(_.*)?$ ]]; then
+                local year="${BASH_REMATCH[1]}"; local month="${BASH_REMATCH[2]}"; local day="${BASH_REMATCH[3]}"; local hour="${BASH_REMATCH[4]}"
+                timestamp="${year}${month}${day}${hour}0000"
+            elif [[ "$period_key" =~ ^([0-9]{4})-([0-9]{2})-([0-9]{2})_([0-9]{2})-([0-9]{2})(_.*)?$ ]]; then
+                local year="${BASH_REMATCH[1]}"; local month="${BASH_REMATCH[2]}"; local day="${BASH_REMATCH[3]}"; local hour="${BASH_REMATCH[4]}"; local minute="${BASH_REMATCH[5]}"
+                timestamp="${year}${month}${day}${hour}${minute}00"
+            elif [[ "$period_key" =~ ^([0-9]{4})-([0-9]{2})-([0-9]{2})_([0-9]{2})-([0-9]{2})-([0-9]{2})_UTC(_.*)?$ ]]; then
+                local year="${BASH_REMATCH[1]}"; local month="${BASH_REMATCH[2]}"; local day="${BASH_REMATCH[3]}"; local hour="${BASH_REMATCH[4]}"; local minute="${BASH_REMATCH[5]}"; local second="${BASH_REMATCH[6]}"
+                timestamp="${year}${month}${day}${hour}${minute}${second}"
+            fi
+            if [ -n "$timestamp" ]; then
+                printf '%s:%s\n' "$timestamp" "$archive_file"
+            fi
+        done
         for archive_file in "$folder"/*.tar.gz; do
             [ -f "$archive_file" ] || continue
             local basename_archive
@@ -55,8 +88,8 @@ find_newest_archive() {
             if [ -n "$timestamp" ]; then
                 printf '%s:%s\n' "$timestamp" "$archive_file"
             fi
-        done | LC_ALL=C sort -r
-    )
+        done
+    ) | LC_ALL=C sort -r
     shopt -u nullglob
 
     if [ -n "$newest_archive" ]; then

archive-cron-logs.sh

@@ -42,7 +42,14 @@ fi
 previous_period=""
 newest_archive=$(find_newest_archive "$CRON_LOG_DIR")
 if [ -n "$newest_archive" ]; then
-    archive_basename=$(basename "$newest_archive" .tar.gz)
+    archive_parent=$(dirname "$newest_archive")
+    archive_base=$(basename "$newest_archive" .tar.gz)
+    if [ "$archive_parent" != "$CRON_LOG_DIR" ]; then
+        year_subdir=$(basename "$archive_parent")
+        archive_basename="${year_subdir}-${archive_base}"
+    else
+        archive_basename="$archive_base"
+    fi
     previous_period=$(parse_previous_period_from_archive_name "$archive_basename")
 fi
 
@@ -110,7 +117,16 @@ while :; do
                 if [ -n "$common_task" ]; then
                     archive_name="${archive_name}_${common_task}"
                 fi
-                archive_path="$CRON_LOG_DIR/${archive_name}.tar.gz"
+                # Group archives by year subdirectory: 2025-01-01_task -> 2025/01-01_task.tar.gz
+                archive_year="${archive_name:0:4}"
+                if [ "${#archive_name}" -gt 4 ]; then
+                    archive_rest="${archive_name:5}"
+                else
+                    archive_rest="$archive_name"
+                fi
+                archive_year_dir="$CRON_LOG_DIR/$archive_year"
+                mkdir -p "$archive_year_dir"
+                archive_path="$archive_year_dir/${archive_rest}.tar.gz"
 
                 rel_files=()
                 while IFS= read -r f; do