diff --git a/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/agent-install-files/amd64/cmp-agent_install_common b/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/agent-install-files/amd64/cmp-agent_install_common
index 42d699c1..52a4e95f 100644
--- a/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/agent-install-files/amd64/cmp-agent_install_common
+++ b/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/agent-install-files/amd64/cmp-agent_install_common
@@ -4,11 +4,35 @@
 
 mkdir -p /cmp-agent/
 
+# Another agent's install on this VM may be running podman (e.g. loading its image) while this
+# runs, and the log agent's service runs podman too. Extracting straight over the binaries leaves
+# them missing or half-written for a moment, so exec fails with "Text file busy". Extract to a
+# staging directory on the same filesystem and rename each file into place: rename swaps the
+# file in one step, so podman is always either the old or the new complete binary.
+STAGE=/cmp-agent/.common-stage_"$1"
+rm -rf "$STAGE"
+mkdir -p "$STAGE"
+
 echo "[*] Extracting agent common files..."
-/tmp/busybox_"$1" tar xvf /tmp/cmp-agent-common_"$1.tar.gz" -C /cmp-agent/
+/tmp/busybox_"$1" tar xvf /tmp/cmp-agent-common_"$1.tar.gz" -C "$STAGE"
 STATUS=`echo $?`
 if [ "$STATUS" != 0 ]; then
   echo "[!] ERROR: Failed to extract agent common files!"
+  rm -rf "$STAGE"
+  exit 1;
+fi
+
+echo "[*] Moving agent common files into place..."
+(cd "$STAGE" && find . -mindepth 1 -type d) | while read -r d; do
+  mkdir -p "/cmp-agent/$d" && chmod --reference="$STAGE/$d" "/cmp-agent/$d"
+done &&
+(cd "$STAGE" && find . ! -type d) | while read -r f; do
+  mv -f "$STAGE/$f" "/cmp-agent/$f" || exit 1
+done
+STATUS=`echo $?`
+rm -rf "$STAGE"
+if [ "$STATUS" != 0 ]; then
+  echo "[!] ERROR: Failed to move agent common files into place!"
   exit 1;
 fi
 
@@ -21,47 +45,76 @@ chcon -Rt bin_t /cmp-agent/common/iptables/bin/
 echo "[*] Fixing permission for '/cmp-agent/common/iptables/sbin/'..."
 chcon -Rt bin_t /cmp-agent/common/iptables/sbin/
 
+# fluent-bit tails /var/log/syslog, which rsyslog writes. Some CSP images (e.g. minimal
+# Ubuntu/Rocky) ship without rsyslog, so install it first. If the package repo is unreachable
+# the install fails; skip the rsyslog steps then so the metric agent still installs - the
+# fluent-bit playbook checks rsyslog itself and fails the log agent with a clear message.
+install_rsyslog() {
+  if command -v apt-get >/dev/null 2>&1; then
+    # A freshly created VM may still hold the apt lock (cloud-init, unattended-upgrades).
+    DEBIAN_FRONTEND=noninteractive timeout 900 apt-get -o DPkg::Lock::Timeout=300 update -q < /dev/null &&
+      DEBIAN_FRONTEND=noninteractive timeout 900 apt-get -o DPkg::Lock::Timeout=300 install -y -q rsyslog < /dev/null
+  elif command -v dnf >/dev/null 2>&1; then
+    timeout 900 dnf install -y -q rsyslog < /dev/null
+  elif command -v yum >/dev/null 2>&1; then
+    timeout 900 yum install -y -q rsyslog < /dev/null
+  elif command -v zypper >/dev/null 2>&1; then
+    timeout 900 zypper --non-interactive install rsyslog < /dev/null
+  else
+    echo "[!] No supported package manager found (apt-get/dnf/yum/zypper)."
+    return 1
+  fi
+}
+
 echo "[*] Getting rsyslog service configuration path..."
+# Without rsyslog, systemctl show prints an empty FragmentPath and still exits 0, so check the
+# value itself. Running the steps below with an empty path makes grep wait on stdin forever.
 RSYSLOG_PATH=`systemctl show rsyslog -p FragmentPath | cut -d= -f2`
-STATUS=`echo $?`
-if [ "$STATUS" != 0 ]; then
-  echo "[!] ERROR: Failed to get rsyslog service configuration path."
-  exit 1;
+if [ -z "$RSYSLOG_PATH" ]; then
+  echo "[*] rsyslog service not found - installing rsyslog..."
+  if install_rsyslog; then
+    systemctl enable --now rsyslog
+    RSYSLOG_PATH=`systemctl show rsyslog -p FragmentPath | cut -d= -f2`
+  fi
 fi
 
-echo "[*] Removing existing TZ environment variable if exists in rsyslog.service..."
-TZ_EXISTS=`grep -c "Environment=TZ" $RSYSLOG_PATH`
-if [ "$TZ_EXISTS" != "" ]; then
-  sed -i '/^Environment=TZ.*/d' $RSYSLOG_PATH | true
-fi
-STATUS=`echo $?`
-if [ "$STATUS" != 0 ]; then
-  echo "[!] ERROR: Failed to remove existing TZ environment variable if exists in rsyslog.service."
-  exit 1;
-fi
+if [ -z "$RSYSLOG_PATH" ]; then
+  echo "[!] rsyslog install failed - skipping rsyslog TZ configuration. Host logs will not be collected."
+else
+  echo "[*] Removing existing TZ environment variable if exists in rsyslog.service..."
+  TZ_EXISTS=`grep -c "Environment=TZ" "$RSYSLOG_PATH"`
+  if [ "$TZ_EXISTS" != "" ]; then
+    sed -i '/^Environment=TZ.*/d' "$RSYSLOG_PATH" | true
+  fi
+  STATUS=`echo $?`
+  if [ "$STATUS" != 0 ]; then
+    echo "[!] ERROR: Failed to remove existing TZ environment variable if exists in rsyslog.service."
+    exit 1;
+  fi
 
-echo "[*] Adding new TZ environment variable to rsyslog.service..."
-sed -i '/^\[Service\]/a Environment=TZ=Asia/Seoul' $RSYSLOG_PATH
-STATUS=`echo $?`
-if [ "$STATUS" != 0 ]; then
-  echo "[!] ERROR: Failed to add new TZ environment variable to rsyslog.service."
-  exit 1;
-fi
+  echo "[*] Adding new TZ environment variable to rsyslog.service..."
+  sed -i '/^\[Service\]/a Environment=TZ=Asia/Seoul' "$RSYSLOG_PATH"
+  STATUS=`echo $?`
+  if [ "$STATUS" != 0 ]; then
+    echo "[!] ERROR: Failed to add new TZ environment variable to rsyslog.service."
+    exit 1;
+  fi
 
-echo "[*] Reloading systemctl daemon..."
-systemctl daemon-reload
-STATUS=`echo $?`
-if [ "$STATUS" != 0 ]; then
-  echo "[!] ERROR: Failed to reload systemctl daemon."
-  exit 1;
-fi
+  echo "[*] Reloading systemctl daemon..."
+  systemctl daemon-reload
+  STATUS=`echo $?`
+  if [ "$STATUS" != 0 ]; then
+    echo "[!] ERROR: Failed to reload systemctl daemon."
+    exit 1;
+  fi
 
-echo "[*] Restarting rsyslog service..."
-systemctl restart rsyslog
-STATUS=`echo $?`
-if [ "$STATUS" != 0 ]; then
-  echo "[!] ERROR: Failed to restart rsyslog service."
-  exit 1;
+  echo "[*] Restarting rsyslog service..."
+  systemctl restart rsyslog
+  STATUS=`echo $?`
+  if [ "$STATUS" != 0 ]; then
+    echo "[!] ERROR: Failed to restart rsyslog service."
+    exit 1;
+  fi
 fi
 
 echo "[*] Creating syslog symbolic link file for RHEL based OS..."
diff --git a/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/agent-install-files/arm64/cmp-agent_install_common b/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/agent-install-files/arm64/cmp-agent_install_common
index 42d699c1..52a4e95f 100644
--- a/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/agent-install-files/arm64/cmp-agent_install_common
+++ b/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/agent-install-files/arm64/cmp-agent_install_common
@@ -4,11 +4,35 @@
 
 mkdir -p /cmp-agent/
 
+# Another agent's install on this VM may be running podman (e.g. loading its image) while this
+# runs, and the log agent's service runs podman too. Extracting straight over the binaries leaves
+# them missing or half-written for a moment, so exec fails with "Text file busy". Extract to a
+# staging directory on the same filesystem and rename each file into place: rename swaps the
+# file in one step, so podman is always either the old or the new complete binary.
+STAGE=/cmp-agent/.common-stage_"$1"
+rm -rf "$STAGE"
+mkdir -p "$STAGE"
+
 echo "[*] Extracting agent common files..."
-/tmp/busybox_"$1" tar xvf /tmp/cmp-agent-common_"$1.tar.gz" -C /cmp-agent/
+/tmp/busybox_"$1" tar xvf /tmp/cmp-agent-common_"$1.tar.gz" -C "$STAGE"
 STATUS=`echo $?`
 if [ "$STATUS" != 0 ]; then
   echo "[!] ERROR: Failed to extract agent common files!"
+  rm -rf "$STAGE"
+  exit 1;
+fi
+
+echo "[*] Moving agent common files into place..."
+(cd "$STAGE" && find . -mindepth 1 -type d) | while read -r d; do
+  mkdir -p "/cmp-agent/$d" && chmod --reference="$STAGE/$d" "/cmp-agent/$d"
+done &&
+(cd "$STAGE" && find . ! -type d) | while read -r f; do
+  mv -f "$STAGE/$f" "/cmp-agent/$f" || exit 1
+done
+STATUS=`echo $?`
+rm -rf "$STAGE"
+if [ "$STATUS" != 0 ]; then
+  echo "[!] ERROR: Failed to move agent common files into place!"
   exit 1;
 fi
 
@@ -21,47 +45,76 @@ chcon -Rt bin_t /cmp-agent/common/iptables/bin/
 echo "[*] Fixing permission for '/cmp-agent/common/iptables/sbin/'..."
 chcon -Rt bin_t /cmp-agent/common/iptables/sbin/
 
+# fluent-bit tails /var/log/syslog, which rsyslog writes. Some CSP images (e.g. minimal
+# Ubuntu/Rocky) ship without rsyslog, so install it first. If the package repo is unreachable
+# the install fails; skip the rsyslog steps then so the metric agent still installs - the
+# fluent-bit playbook checks rsyslog itself and fails the log agent with a clear message.
+install_rsyslog() {
+  if command -v apt-get >/dev/null 2>&1; then
+    # A freshly created VM may still hold the apt lock (cloud-init, unattended-upgrades).
+    DEBIAN_FRONTEND=noninteractive timeout 900 apt-get -o DPkg::Lock::Timeout=300 update -q < /dev/null &&
+      DEBIAN_FRONTEND=noninteractive timeout 900 apt-get -o DPkg::Lock::Timeout=300 install -y -q rsyslog < /dev/null
+  elif command -v dnf >/dev/null 2>&1; then
+    timeout 900 dnf install -y -q rsyslog < /dev/null
+  elif command -v yum >/dev/null 2>&1; then
+    timeout 900 yum install -y -q rsyslog < /dev/null
+  elif command -v zypper >/dev/null 2>&1; then
+    timeout 900 zypper --non-interactive install rsyslog < /dev/null
+  else
+    echo "[!] No supported package manager found (apt-get/dnf/yum/zypper)."
+    return 1
+  fi
+}
+
 echo "[*] Getting rsyslog service configuration path..."
+# Without rsyslog, systemctl show prints an empty FragmentPath and still exits 0, so check the
+# value itself. Running the steps below with an empty path makes grep wait on stdin forever.
 RSYSLOG_PATH=`systemctl show rsyslog -p FragmentPath | cut -d= -f2`
-STATUS=`echo $?`
-if [ "$STATUS" != 0 ]; then
-  echo "[!] ERROR: Failed to get rsyslog service configuration path."
-  exit 1;
+if [ -z "$RSYSLOG_PATH" ]; then
+  echo "[*] rsyslog service not found - installing rsyslog..."
+  if install_rsyslog; then
+    systemctl enable --now rsyslog
+    RSYSLOG_PATH=`systemctl show rsyslog -p FragmentPath | cut -d= -f2`
+  fi
 fi
 
-echo "[*] Removing existing TZ environment variable if exists in rsyslog.service..."
-TZ_EXISTS=`grep -c "Environment=TZ" $RSYSLOG_PATH`
-if [ "$TZ_EXISTS" != "" ]; then
-  sed -i '/^Environment=TZ.*/d' $RSYSLOG_PATH | true
-fi
-STATUS=`echo $?`
-if [ "$STATUS" != 0 ]; then
-  echo "[!] ERROR: Failed to remove existing TZ environment variable if exists in rsyslog.service."
-  exit 1;
-fi
+if [ -z "$RSYSLOG_PATH" ]; then
+  echo "[!] rsyslog install failed - skipping rsyslog TZ configuration. Host logs will not be collected."
+else
+  echo "[*] Removing existing TZ environment variable if exists in rsyslog.service..."
+  TZ_EXISTS=`grep -c "Environment=TZ" "$RSYSLOG_PATH"`
+  if [ "$TZ_EXISTS" != "" ]; then
+    sed -i '/^Environment=TZ.*/d' "$RSYSLOG_PATH" | true
+  fi
+  STATUS=`echo $?`
+  if [ "$STATUS" != 0 ]; then
+    echo "[!] ERROR: Failed to remove existing TZ environment variable if exists in rsyslog.service."
+    exit 1;
+  fi
 
-echo "[*] Adding new TZ environment variable to rsyslog.service..."
-sed -i '/^\[Service\]/a Environment=TZ=Asia/Seoul' $RSYSLOG_PATH
-STATUS=`echo $?`
-if [ "$STATUS" != 0 ]; then
-  echo "[!] ERROR: Failed to add new TZ environment variable to rsyslog.service."
-  exit 1;
-fi
+  echo "[*] Adding new TZ environment variable to rsyslog.service..."
+  sed -i '/^\[Service\]/a Environment=TZ=Asia/Seoul' "$RSYSLOG_PATH"
+  STATUS=`echo $?`
+  if [ "$STATUS" != 0 ]; then
+    echo "[!] ERROR: Failed to add new TZ environment variable to rsyslog.service."
+    exit 1;
+  fi
 
-echo "[*] Reloading systemctl daemon..."
-systemctl daemon-reload
-STATUS=`echo $?`
-if [ "$STATUS" != 0 ]; then
-  echo "[!] ERROR: Failed to reload systemctl daemon."
-  exit 1;
-fi
+  echo "[*] Reloading systemctl daemon..."
+  systemctl daemon-reload
+  STATUS=`echo $?`
+  if [ "$STATUS" != 0 ]; then
+    echo "[!] ERROR: Failed to reload systemctl daemon."
+    exit 1;
+  fi
 
-echo "[*] Restarting rsyslog service..."
-systemctl restart rsyslog
-STATUS=`echo $?`
-if [ "$STATUS" != 0 ]; then
-  echo "[!] ERROR: Failed to restart rsyslog service."
-  exit 1;
+  echo "[*] Restarting rsyslog service..."
+  systemctl restart rsyslog
+  STATUS=`echo $?`
+  if [ "$STATUS" != 0 ]; then
+    echo "[!] ERROR: Failed to restart rsyslog service."
+    exit 1;
+  fi
 fi
 
 echo "[*] Creating syslog symbolic link file for RHEL based OS..."
diff --git a/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/tasks/agent_common_begin.yaml b/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/tasks/agent_common_begin.yaml
index a82118ac..fb2348b5 100644
--- a/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/tasks/agent_common_begin.yaml
+++ b/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/tasks/agent_common_begin.yaml
@@ -61,7 +61,9 @@
     ansible_ssh_private_key_file: "{{ target_sshkey_path }}"
   raw: "systemctl status --no-pager rsyslog"
   register: rsyslog_status
-  failed_when: rsyslog_status.rc not in [0, 3]
+  # rc 4 = no rsyslog unit (the install script could not install it). The log agent install
+  # checks for rsyslog itself; don't fail the metric agent install over it.
+  failed_when: rsyslog_status.rc not in [0, 3, 4]
 
 - name: Display rsyslog status
   debug:
diff --git a/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/tasks/agent_fluent-bit_install.yaml b/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/tasks/agent_fluent-bit_install.yaml
index 070b6dbe..cad34ebf 100644
--- a/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/tasks/agent_fluent-bit_install.yaml
+++ b/conf/docker/conf/mc-observability/mc-observability-infra/ansible/playbooks/install-agent/roles/agent/tasks/agent_fluent-bit_install.yaml
@@ -8,6 +8,29 @@
 - name: Begin Agent Common Tasks
   include_tasks: agent_common_begin.yaml
 
+# fluent-bit reads host logs from /var/log/syslog, which rsyslog writes. The common install
+# script tries to install rsyslog; if that failed (e.g. no access to the package repo), fail
+# the log agent here so it shows as FAILED instead of installing and collecting nothing.
+- name: Check rsyslog is running for Fluent Bit
+  become: true
+  become_method: sudo
+  become_user: root
+  vars:
+    ansible_connection: ssh
+    ansible_host: "{{ target_host }}"
+    ansible_port: "{{ target_port }}"
+    ansible_user: "{{ target_user }}"
+    ansible_ssh_private_key_file: "{{ target_sshkey_path }}"
+  raw: "systemctl is-active rsyslog"
+  register: rsyslog_active
+  changed_when: false
+  failed_when: false
+
+- name: Fail when rsyslog is not running
+  fail:
+    msg: "rsyslog is not running on this VM and could not be installed (check that the VM can reach its OS package repository). Fluent Bit collects host logs from /var/log/syslog written by rsyslog."
+  when: rsyslog_active.stdout | trim != 'active'
+
 - name: Compress agent files
   raw: "cd {{ role_path }}/agent-install-files/{{ arch }} && sudo rm -f cmp-agent/cmp-fluent-bit_{{ request_id }}.tar.gz && sudo chown -R root:root fluent-bit/ && sudo ../amd64/busybox tar cvfz cmp-agent/cmp-fluent-bit_{{ request_id }}.tar.gz fluent-bit/ && sudo chmod 644 cmp-agent/cmp-fluent-bit_{{ request_id }}.tar.gz"
   delegate_to: 127.0.0.1
diff --git a/conf/docker/docker-compose.yaml b/conf/docker/docker-compose.yaml
index 567b814a..68d129e0 100644
--- a/conf/docker/docker-compose.yaml
+++ b/conf/docker/docker-compose.yaml
@@ -1254,7 +1254,7 @@ services:
       <<: *default-health-check
 ##### MC-OBSERVABILITY #########################################################################################################################
   mc-observability-manager:
-    image: cloudbaristaorg/mc-observability:0.7.0
+    image: cloudbaristaorg/mc-observability:0.7.1
     container_name: mc-observability-manager
     restart: unless-stopped
     networks:
@@ -1263,6 +1263,10 @@ services:
       - mc-infra-manager-network
       - mc-infra-connector-network
       - mc-web-console-network
+    ports:
+      - target: 18080
+        published: ${MC_OBSERVABILITY_MANAGER_PORT}
+        protocol: tcp
     depends_on:
       mc-observability-grafana:
         condition: service_healthy
@@ -1294,7 +1298,7 @@ services:
       start_period: 5m
 
   mc-observability-front:
-    image: cloudbaristaorg/mc-observability-front:0.7.0
+    image: cloudbaristaorg/mc-observability-front:0.7.1
     container_name: mc-observability-front
     restart: unless-stopped
     networks:
@@ -1306,7 +1310,7 @@ services:
         condition: service_started
 
   mc-observability-infra:
-    image: cloudbaristaorg/mc-observability-infra:0.7.0
+    image: cloudbaristaorg/mc-observability-infra:0.7.1
     container_name: mc-observability-infra
     restart: unless-stopped
     networks:
@@ -1431,7 +1435,7 @@ services:
     init: true
 
   mc-observability-influx:
-    image: cloudbaristaorg/mc-observability-influx:0.7.0
+    image: cloudbaristaorg/mc-observability-influx:0.7.1
     container_name: mc-observability-influx
     restart: unless-stopped
     networks:
@@ -1473,7 +1477,7 @@ services:
       - mc-observability-network
 
   mc-observability-influx-2:
-    image: cloudbaristaorg/mc-observability-influx:0.7.0
+    image: cloudbaristaorg/mc-observability-influx:0.7.1
     container_name: mc-observability-influx-2
     restart: unless-stopped
     networks:
@@ -1649,7 +1653,7 @@ services:
     init: true
 
   mc-observability-grafana:
-    image: cloudbaristaorg/mc-observability-grafana:0.7.0
+    image: cloudbaristaorg/mc-observability-grafana:0.7.1
     container_name: mc-observability-grafana
     restart: unless-stopped
     networks:
@@ -1684,7 +1688,7 @@ services:
         fi
 
   mc-observability-insight:
-    image: cloudbaristaorg/mc-observability-insight:0.7.0
+    image: cloudbaristaorg/mc-observability-insight:0.7.1
     container_name: mc-observability-insight
     restart: always
     networks:
@@ -1713,7 +1717,7 @@ services:
       start_period: 10s
 
   mc-observability-insight-scheduler:
-    image: cloudbaristaorg/mc-observability-insight-scheduler:0.7.0
+    image: cloudbaristaorg/mc-observability-insight-scheduler:0.7.1
     container_name: mc-observability-insight-scheduler
     restart: always
     ports:
@@ -1801,7 +1805,7 @@ services:
       - mc-observability-network
 
   mc-observability-mcp-maria:
-    image: cloudbaristaorg/mc-observability-mcp-mariadb:0.7.0
+    image: cloudbaristaorg/mc-observability-mcp-mariadb:0.7.1
     container_name: mc-observability-mcp-maria
     restart: always
     networks:
@@ -1825,7 +1829,7 @@ services:
       - OPENAI_API_KEY=${MC_OBSERVABILITY_OPENAI_API_KEY:-}
 
   mc-observability-mcp-influx:
-    image: cloudbaristaorg/mc-observability-mcp-influxdb:0.7.0
+    image: cloudbaristaorg/mc-observability-mcp-influxdb:0.7.1
     container_name: mc-observability-mcp-influx
     restart: always
     ports: