Run obfi_previous_minute inside docker to prevent orphaned tac command

Author: cdrini

scripts/monitoring/tests/test_utils_sh.py

@@ -34,10 +34,10 @@ def test_log_recent_bot_traffic():
         # alias the obfi commands to return some noise
         aliases_fp.write(
             f"""
-                obfi_previous_minute() {{
+                obfi_in_docker() {{
                     cat scripts/monitoring/tests/sample_covers_nginx_logs.log
                 }}
-                export -f obfi_previous_minute
+                export -f obfi_in_docker
 
                 nc() {{
                     # Read stdin and write to nc_fp
@@ -62,7 +62,7 @@ def test_log_recent_bot_traffic():
             expected_output = """
 stats.ol-covers0.bot_traffic.gptbot 2 1741054377
 stats.ol-covers0.bot_traffic.meta_externalagent 1 1741054377
-stats.ol-covers0.bot_traffic.other 0 1741054377
+stats.ol-covers0.bot_traffic.other 1 1741054377
 stats.ol-covers0.bot_traffic.non_bot 6 1741054377
             """.strip()
             assert f.read().strip() == expected_output
@@ -76,10 +76,10 @@ def test_log_recent_http_statuses():
         # alias the obfi commands to return some noise
         aliases_fp.write(
             f"""
-                obfi_previous_minute() {{
+                obfi_in_docker() {{
                     cat scripts/monitoring/tests/sample_covers_nginx_logs.log
                 }}
-                export -f obfi_previous_minute
+                export -f obfi_in_docker
 
                 nc() {{
                     # Read stdin and write to nc_fp

scripts/monitoring/utils.sh

@@ -75,7 +75,7 @@ log_recent_bot_traffic() {
     BUCKET="$1"
 
     # Get top bot user agent counts in the last minute
-    BOT_TRAFFIC_COUNTS=$(obfi_previous_minute | obfi_top_bots)
+    BOT_TRAFFIC_COUNTS=$(obfi_in_docker obfi_previous_minute | obfi_top_bots)
 
     # Output like this:
     # 1412516 bingbot
@@ -101,9 +101,9 @@ log_recent_bot_traffic() {
 
     # Also log other bots as a single metric
     OTHER_BOTS_COUNT=$(
-        obfi_previous_minute | \
+        obfi_in_docker obfi_previous_minute | \
         grep -iE '\b[a-z_-]+(bot|spider|crawler)' | \
-        grep -viE 'bingbot|claudebot|googlebot|applebot|gptbot|yandex.com/bots|ahrefsbot|amazonbot|petalbot|brightbot|SemanticScholarBot|uptimerobot|seznamebot|OAI-SearchBot|VsuSearchSpider' | \
+        obfi_top_bots -v | \
         wc -l
     )
 
@@ -113,7 +113,7 @@ log_recent_bot_traffic() {
 
     # And finally, also log non bot traffic
     NON_BOT_TRAFFIC_COUNT=$(
-        obfi_previous_minute | \
+        obfi_in_docker obfi_previous_minute | \
         grep -viE '\b[a-z_-]+(bot|spider|crawler)' | \
         wc -l
     )
@@ -128,7 +128,7 @@ log_recent_http_statuses() {
     BUCKET="$1"
 
     # Get top http counts from previous minute
-    HTTP_STATUS_COUNTS=$(obfi_previous_minute | obfi_top_http_statuses)
+    HTTP_STATUS_COUNTS=$(obfi_in_docker obfi_previous_minute | obfi_top_http_statuses)
     # Output like this:
     #   60319 " 200
     #   55926 " 302
@@ -157,7 +157,7 @@ log_top_ip_counts() {
 
     # Get top IP counts in the last minute
     TOP_IP_COUNTS=$(
-        obfi_previous_minute | \
+        obfi_in_docker obfi_previous_minute | \
         obfi_top_ips 25 | \
         awk '{print $1}'
     )

scripts/obfi.sh

@@ -64,13 +64,18 @@ obfi() {
         return 0
     fi
     CMD=${1:-"tail -f"}
+    CONTAINER="${CONTAINER:-}"
+    OBFI_SUDO=${OBFI_SUDO-"sudo"}
+    local sudo_cmd=""
+
+    if [ ! -z "$OBFI_SUDO" ]; then
+        sudo_cmd="$OBFI_SUDO -E"
+    fi
 
     if command -v docker >/dev/null 2>&1; then
         if [ -z "$CONTAINER" ]; then
             CONTAINER=$(docker ps --format '{{.Names}}' | grep nginx)
         fi
-    else
-        CONTAINER=""
     fi
 
     if [ ! -z "$CONTAINER" ]; then
@@ -82,8 +87,34 @@ obfi() {
         LOG_DIR="${OBFI_LOG_DIR:-/1/var/log/nginx}"
         FILE=${2:-"$LOG_DIR/access.log"}
         echo "Reading from $FILE" 1>&2
-        sudo -E $CMD $FILE
+        $sudo_cmd $CMD $FILE
+    fi
+}
+
+obfi_in_docker() {
+    if ! command -v docker >/dev/null 2>&1; then
+        echo "Docker is not installed" 1>&2
+        return 1
+    fi
+
+    CMD=${1:-"tail -f"}
+    CONTAINER="${CONTAINER:-}"
+
+    if [ -z "$CONTAINER" ]; then
+        CONTAINER=$(docker ps --format '{{.Names}}' | grep nginx)
+    fi
+
+    if [ -z "$CONTAINER" ]; then
+        echo "No nginx container found"
+        return 1
     fi
+
+    docker exec "$CONTAINER" bash -c "
+        source /openlibrary/scripts/obfi.sh
+        export OBFI_SUDO=''
+        export OBFI_LOG_DIR='/var/log/nginx'
+        $CMD
+    "
 }
 
 ###############################################################