diff --git a/README.rst b/README.rst
index 33a3bf4..355c68f 100644
--- a/README.rst
+++ b/README.rst
@@ -21,6 +21,7 @@
         metric_collector_host: 127.0.0.1
         metric_collector_port: 5567
         poolsize: 100
+        max_message_size: 262144
 
 Default values:
 
@@ -30,6 +31,7 @@
 * ``metric_collector_host: 127.0.0.1``
 * ``metric_collector_port: 5567``
 * ``poolsize: 100``
+* ``max_message_size: 262144``
 
 Local Metric collector service
 
@@ -54,6 +56,7 @@
         nagios_port: 5601
         nagios_username: nagiosadmin
         poolsize: 100
+        max_message_size: 262144
 
 Default values:
 
@@ -65,6 +68,7 @@
 * ``influxdb_timeout: 5000``
 * ``nagios_port: 8001``
 * ``poolsize: 100``
+* ``max_message_size: 262144``
 
 Remote Metric Collector service
 
@@ -91,6 +95,7 @@
         influxdb_time_precision: ms
         influxdb_username: lma
         poolsize: 100
+        max_message_size: 262144
 
 Default values:
 
@@ -105,6 +110,7 @@
 * ``influxdb_time_precision: ms``
 * ``influxdb_timeout: 5000``
 * ``poolsize: 100``
+* ``max_message_size: 262144``
 
 Aggregator service
 
@@ -127,6 +133,7 @@
         nagios_port: 5601
         nagios_username: nagiosadmin
         poolsize: 100
+        max_message_size: 262144
 
 Default values:
 
@@ -138,6 +145,7 @@
 * ``nagios_port: 8001``
 * ``nagios_default_host_alarm_clusters: 00-clusters``
 * ``poolsize: 100``
+* ``max_message_size: 262144``
 
 Ceilometer service
 
diff --git a/heka/_common.sls b/heka/_common.sls
index a46b820..fc1a651 100644
--- a/heka/_common.sls
+++ b/heka/_common.sls
@@ -26,6 +26,11 @@
   - require:
     - user: heka_user
 
+/usr/local/bin/monitor_heka_queues.sh:
+  file.managed:
+  - source: salt://heka/files/monitor_heka_queues.sh
+  - mode: 755
+
 heka_user:
   user.present:
   - name: heka
diff --git a/heka/_service.sls b/heka/_service.sls
index 7af3101..c421697 100644
--- a/heka/_service.sls
+++ b/heka/_service.sls
@@ -220,6 +220,7 @@
   - defaults:
     service_name: {{ service_name }}
     poolsize: {{ server.poolsize }}
+    max_message_size: {{ server.max_message_size }}
   - require:
     - file: heka_{{ service_name }}_conf_dir
   - require_in:
diff --git a/heka/files/monitor_heka_queues.sh b/heka/files/monitor_heka_queues.sh
new file mode 100755
index 0000000..7360440
--- /dev/null
+++ b/heka/files/monitor_heka_queues.sh
@@ -0,0 +1,13 @@
+#!/bin/bash -e
+#
+# This script reports the size of each Heka output queue.
+
+for q in /var/cache/*/output_queue/*;
+do
+  DATA=$(du -sb "$q")
+  SIZE=$(echo $DATA | cut -d " " -f 1)
+  QUEUE=$(echo $DATA | cut -d " " -f 2)
+  if [[ -n "${SIZE}" && -n "${QUEUE}" ]]; then
+    echo "heka_output_queue_size,queue=${QUEUE} value=${SIZE}"
+  fi
+done
diff --git a/heka/files/toml/global.toml b/heka/files/toml/global.toml
index 30d460d..bdc2a81 100644
--- a/heka/files/toml/global.toml
+++ b/heka/files/toml/global.toml
@@ -6,7 +6,7 @@
 
 hostname="{{ grains.fqdn.split('.')[0] }}"
 
-max_message_size = 262144
+max_message_size = {{ max_message_size }}
 max_process_inject = 1
 max_timer_inject = 10
 poolsize = {{ poolsize }}
diff --git a/heka/map.jinja b/heka/map.jinja
index 297705f..07e39a2 100644
--- a/heka/map.jinja
+++ b/heka/map.jinja
@@ -54,6 +54,7 @@
     'emit_rates': True,
     'prefix_dir': default_prefix_dir,
     'elasticsearch_port': default_elasticsearch_port,
+    'max_message_size': 262144,
     'poolsize': 100,
     'automatic_starting': default_automatic_starting,
     'metric_collector_host': '127.0.0.1',
@@ -76,6 +77,7 @@
     'influxdb_tag_fields': default_influxdb_tag_fields,
     'aggregator_port': default_aggregator_port,
     'nagios_port': default_nagios_port,
+    'max_message_size': 262144,
     'poolsize': 100,
     'automatic_starting': default_automatic_starting,
     'sensu_noop_handler': default_sensu_noop_handler,
@@ -99,6 +101,7 @@
     'influxdb_timeout': default_influxdb_timeout,
     'influxdb_tag_fields': default_influxdb_tag_fields,
     'aggregator_port': default_aggregator_port,
+    'max_message_size': 262144,
     'poolsize': 100,
     'automatic_starting': default_automatic_starting,
     'sensu_noop_handler': default_sensu_noop_handler,
@@ -119,6 +122,7 @@
     'influxdb_tag_fields': default_influxdb_tag_fields,
     'nagios_port': default_nagios_port,
     'nagios_default_host_alarm_clusters': default_nagios_host_alarm_clusters,
+    'max_message_size': 262144,
     'poolsize': 100,
     'automatic_starting': default_automatic_starting,
     'sensu_noop_handler': default_sensu_noop_handler,
@@ -141,6 +145,7 @@
     'amqp_vhost': '/openstack',
     'amqp_queue': 'metering.sample',
     'resource_decoding': False,
+    'max_message_size': 262144,
     'poolsize': 100,
     'automatic_starting': default_automatic_starting,
     'sensu_noop_handler': default_sensu_noop_handler,
diff --git a/heka/meta/prometheus.yml b/heka/meta/prometheus.yml
new file mode 100644
index 0000000..daf3bd8
--- /dev/null
+++ b/heka/meta/prometheus.yml
@@ -0,0 +1,13 @@
+server:
+  alert:
+    HekaOutputQueueStalled:
+      if: >-
+        delta(heka_output_queue_size[1h]) == 0
+      labels:
+        severity: warning
+        service: heka
+      annotations:
+        summary: 'Heka queue stalled'
+{%- raw %}
+        description: 'The {{ $labels.queue }} queue is stalled on node {{ $labels.host }} for more than 1 hour. The corresponding Heka service is either down or stuck.'
+{%- endraw %}
diff --git a/heka/meta/sphinx.yml b/heka/meta/sphinx.yml
index 176da9a..d971b22 100644
--- a/heka/meta/sphinx.yml
+++ b/heka/meta/sphinx.yml
@@ -8,7 +8,7 @@
       param:
         version:
           name: "Version"
-          value: "{{ salt['cmd.run']('hekad --version 2>/dev/null || echo unknown') }}"
+          value: "{{ salt['cmd.shell']('hekad --version 2>/dev/null || echo unknown') }}"
   {%- endif %}
   {%- if pillar.heka.log_collector is defined %}
     log_collector:
@@ -16,7 +16,7 @@
       param:
         version:
           name: "Version"
-          value: "{{ salt['cmd.run']('hekad --version 2>/dev/null || echo unknown') }}"
+          value: "{{ salt['cmd.shell']('hekad --version 2>/dev/null || echo unknown') }}"
   {%- endif %}
   {%- if pillar.heka.metric_collector is defined %}
     metric_collector:
@@ -24,7 +24,7 @@
       param:
         version:
           name: "Version"
-          value: "{{ salt['cmd.run']('hekad --version 2>/dev/null || echo unknown') }}"
+          value: "{{ salt['cmd.shell']('hekad --version 2>/dev/null || echo unknown') }}"
   {%- endif %}
   {%- if pillar.heka.aggregator_collector is defined %}
     aggregator:
@@ -32,5 +32,5 @@
       param:
         version:
           name: "Version"
-          value: "{{ salt['cmd.run']('hekad --version 2>/dev/null || echo unknown') }}"
+          value: "{{ salt['cmd.shell']('hekad --version 2>/dev/null || echo unknown') }}"
   {%- endif %}
diff --git a/heka/meta/telegraf.yml b/heka/meta/telegraf.yml
new file mode 100644
index 0000000..d8c7b13
--- /dev/null
+++ b/heka/meta/telegraf.yml
@@ -0,0 +1,8 @@
+agent:
+  input:
+    monitor_heka:
+      template: telegraf/files/input/exec.conf
+      commands:
+      - /usr/local/bin/monitor_heka_queues.sh
+      data_format: influx
+      interval: 60s
diff --git a/metadata/service/aggregator/cluster.yml b/metadata/service/aggregator/cluster.yml
index e95243e..35e854b 100644
--- a/metadata/service/aggregator/cluster.yml
+++ b/metadata/service/aggregator/cluster.yml
@@ -5,9 +5,11 @@
 parameters:
   _param:
     aggregator_poolsize: 100
+    aggregator_max_message_size: 262144
   heka:
     aggregator:
       automatic_starting: false
       enabled: true
       influxdb_time_precision: ms
       poolsize: ${_param:aggregator_poolsize}
+      max_message_size: ${_param:aggregator_max_message_size}
\ No newline at end of file
diff --git a/metadata/service/aggregator/single.yml b/metadata/service/aggregator/single.yml
index bf2364e..dc53ba8 100644
--- a/metadata/service/aggregator/single.yml
+++ b/metadata/service/aggregator/single.yml
@@ -5,9 +5,11 @@
 parameters:
   _param:
     aggregator_poolsize: 100
+    aggregator_max_message_size: 262144
   heka:
     aggregator:
       automatic_starting: true
       enabled: true
       influxdb_time_precision: ms
-      poolsize: ${_param:aggregator_poolsize}
\ No newline at end of file
+      poolsize: ${_param:aggregator_poolsize}
+      max_message_size: ${_param:aggregator_max_message_size}
\ No newline at end of file
diff --git a/metadata/service/ceilometer_collector/single.yml b/metadata/service/ceilometer_collector/single.yml
index eacb11d..16f8050 100644
--- a/metadata/service/ceilometer_collector/single.yml
+++ b/metadata/service/ceilometer_collector/single.yml
@@ -5,9 +5,11 @@
 parameters:
   _param:
     ceilometer_collector_poolsize: 100
+    ceilometer_collector_max_message_size: 262144
   heka:
     ceilometer_collector:
       automatic_starting: true
       enabled: true
       influxdb_time_precision: ms
       poolsize: ${_param:ceilometer_collector_poolsize}
+      max_message_size: ${_param:ceilometer_collector_max_message_size}
diff --git a/metadata/service/log_collector/single.yml b/metadata/service/log_collector/single.yml
index 5160c35..d7a53fd 100644
--- a/metadata/service/log_collector/single.yml
+++ b/metadata/service/log_collector/single.yml
@@ -5,7 +5,9 @@
 parameters:
   _param:
     log_collector_poolsize: 100
+    log_collector_max_message_size: 262144
   heka:
     log_collector:
       enabled: true
       poolsize: ${_param:log_collector_poolsize}
+      max_message_size: ${_param:log_collector_max_message_size}
diff --git a/metadata/service/metric_collector/single.yml b/metadata/service/metric_collector/single.yml
index ad183a2..420f630 100644
--- a/metadata/service/metric_collector/single.yml
+++ b/metadata/service/metric_collector/single.yml
@@ -5,8 +5,10 @@
 parameters:
   _param:
     metric_collector_poolsize: 100
+    metric_collector_max_message_size: 262144
   heka:
     metric_collector:
       enabled: true
       influxdb_time_precision: ms
       poolsize: ${_param:metric_collector_poolsize}
+      max_message_size: ${_param:metric_collector_max_message_size}
diff --git a/metadata/service/remote_collector/cluster.yml b/metadata/service/remote_collector/cluster.yml
index 9e60e86..daa7e64 100644
--- a/metadata/service/remote_collector/cluster.yml
+++ b/metadata/service/remote_collector/cluster.yml
@@ -5,9 +5,11 @@
 parameters:
   _param:
     remote_collector_poolsize: 100
+    remote_collector_max_message_size: 262144
   heka:
     remote_collector:
       automatic_starting: false
       enabled: true
       influxdb_time_precision: ms
       poolsize: ${_param:remote_collector_poolsize}
+      max_message_size: ${_param:remote_collector_max_message_size}
diff --git a/metadata/service/remote_collector/single.yml b/metadata/service/remote_collector/single.yml
index 8533ad2..acc2853 100644
--- a/metadata/service/remote_collector/single.yml
+++ b/metadata/service/remote_collector/single.yml
@@ -5,9 +5,11 @@
 parameters:
   _param:
     remote_collector_poolsize: 100
+    remote_collector_max_message_size: 262144
   heka:
     remote_collector:
       automatic_starting: true
       enabled: true
       influxdb_time_precision: ms
       poolsize: ${_param:remote_collector_poolsize}
+      max_message_size: ${_param:remote_collector_max_message_size}
diff --git a/metadata/service/support.yml b/metadata/service/support.yml
index 14a0a96..15b34cb 100644
--- a/metadata/service/support.yml
+++ b/metadata/service/support.yml
@@ -5,7 +5,11 @@
         enabled: true
       heka:
         enabled: true
+      prometheus:
+        enabled: true
       sensu:
         enabled: true
       sphinx:
         enabled: true
+      telegraf:
+        enabled: true
diff --git a/tests/run_tests.sh b/tests/run_tests.sh
index 3f42101..29fb975 100755
--- a/tests/run_tests.sh
+++ b/tests/run_tests.sh
@@ -6,11 +6,13 @@
 CURDIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
 METADATA=${CURDIR}/../metadata.yml
 FORMULA_NAME=$(cat $METADATA | python -c "import sys,yaml; print yaml.load(sys.stdin)['name']")
+FORMULA_META_DIR=${CURDIR}/../${FORMULA_NAME}/meta
 
 ## Overrideable parameters
 PILLARDIR=${PILLARDIR:-${CURDIR}/pillar}
 BUILDDIR=${BUILDDIR:-${CURDIR}/build}
 VENV_DIR=${VENV_DIR:-${BUILDDIR}/virtualenv}
+MOCK_BIN_DIR=${MOCK_BIN_DIR:-${CURDIR}/mock_bin}
 DEPSDIR=${BUILDDIR}/deps
 
 SALT_FILE_DIR=${SALT_FILE_DIR:-${BUILDDIR}/file_root}
@@ -18,7 +20,7 @@
 SALT_CONFIG_DIR=${SALT_CONFIG_DIR:-${BUILDDIR}/salt}
 SALT_CACHE_DIR=${SALT_CACHE_DIR:-${SALT_CONFIG_DIR}/cache}
 
-SALT_OPTS="${SALT_OPTS} --retcode-passthrough --local -c ${SALT_CONFIG_DIR}"
+SALT_OPTS="${SALT_OPTS} --retcode-passthrough --local -c ${SALT_CONFIG_DIR} --log-file=/dev/null"
 
 if [ "x${SALT_VERSION}" != "x" ]; then
     PIP_SALT_VERSION="==${SALT_VERSION}"
@@ -40,10 +42,20 @@
     pip install salt${PIP_SALT_VERSION}
 }
 
+setup_mock_bin() {
+    # If some state requires a binary, a lightweight replacement for
+    # such binary can be put into MOCK_BIN_DIR for test purposes
+    if [ -d "${MOCK_BIN_DIR}" ]; then
+        PATH="${MOCK_BIN_DIR}:$PATH"
+        export PATH
+    fi
+}
+
 setup_pillar() {
     [ ! -d ${SALT_PILLAR_DIR} ] && mkdir -p ${SALT_PILLAR_DIR}
     echo "base:" > ${SALT_PILLAR_DIR}/top.sls
     for pillar in ${PILLARDIR}/*; do
+        grep ${FORMULA_NAME}: ${pillar} &>/dev/null || continue
         state_name=$(basename ${pillar%.sls})
         echo -e "  ${state_name}:\n    - ${state_name}" >> ${SALT_PILLAR_DIR}/top.sls
     done
@@ -56,6 +68,7 @@
 
     echo "base:" > ${SALT_FILE_DIR}/top.sls
     for pillar in ${PILLARDIR}/*.sls; do
+        grep ${FORMULA_NAME}: ${pillar} &>/dev/null || continue
         state_name=$(basename ${pillar%.sls})
         echo -e "  ${state_name}:\n    - ${FORMULA_NAME}" >> ${SALT_FILE_DIR}/top.sls
     done
@@ -64,6 +77,7 @@
 file_client: local
 cachedir: ${SALT_CACHE_DIR}
 verify_env: False
+minion_id_caching: False
 
 file_roots:
   base:
@@ -118,6 +132,7 @@
     [ -d ${BUILDDIR} ] && mkdir -p ${BUILDDIR}
 
     which salt-call || setup_virtualenv
+    setup_mock_bin
     setup_pillar
     setup_salt
     install_dependencies
@@ -125,8 +140,28 @@
 
 run() {
     for pillar in ${PILLARDIR}/*.sls; do
+        grep ${FORMULA_NAME}: ${pillar} &>/dev/null || continue
         state_name=$(basename ${pillar%.sls})
+        salt_run grains.set 'noservices' False force=True
+
+        echo "Checking state ${FORMULA_NAME}.${state_name} ..."
         salt_run --id=${state_name} state.show_sls ${FORMULA_NAME} || (log_err "Execution of ${FORMULA_NAME}.${state_name} failed"; exit 1)
+
+        # Check that all files in 'meta' folder can be rendered using any valid pillar
+        for meta in `find ${FORMULA_META_DIR} -type f`; do
+            meta_name=$(basename ${meta})
+            echo "Checking meta ${meta_name} ..."
+            salt_run --out=quiet --id=${state_name} cp.get_template ${meta} ${SALT_CACHE_DIR}/${meta_name} \
+              || (log_err "Failed to render meta ${meta} using pillar ${FORMULA_NAME}.${state_name}"; exit 1)
+            cat ${SALT_CACHE_DIR}/${meta_name}
+        done
+    done
+}
+
+real_run() {
+    for pillar in ${PILLARDIR}/*.sls; do
+        state_name=$(basename ${pillar%.sls})
+        salt_run --id=${state_name} state.sls ${FORMULA_NAME} || (log_err "Execution of ${FORMULA_NAME}.${state_name} failed"; exit 1)
     done
 }
 
@@ -155,6 +190,9 @@
     run)
         run
         ;;
+    real-run)
+        real_run
+        ;;
     *)
         prepare
         run