Various fixes/changes to our celery setup

[wlach]

@jezdez -- would you mind sanity checking this simple set of fixes when you have a moment?

[jezdez]

Oh one thing may be useful:

diff --git a/bin/run b/bin/run
index 25a6ba3..09b81e9 100755
--- a/bin/run
+++ b/bin/run
@@ -6,6 +6,7 @@ set -eo pipefail
 : "${SLEEP:=1}"
 : "${TRIES:=60}"
 : "${MONITOR_PIDFILE:=/app/celerymonitor.pid}"
+: "${SCHEDULER_PIDFILE:=/app/celerybeat.pid}"

 usage() {
   echo "usage: bin/run web|web-dev|test"
@@ -42,12 +43,13 @@ case $1 in
     exec python manage.py runserver 0.0.0.0:${PORT}
     ;;
   worker)
-    exec newrelic-admin run-program celery -A missioncontrol.celery:celery worker -l info -O fair --events --concurrency=8
+    exec newrelic-admin run-program celery -A missioncontrol.celery:celery worker -l info --events --concurrency=8
     ;;
   scheduler)
+    [ -f ${SCHEDULER_PIDFILE} ] && rm ${SCHEDULER_PIDFILE}
     python manage.py migrate --noinput
     exec newrelic-admin run-program celery -A missioncontrol.celery:celery \
-      beat -l info --pidfile /tmp/celerybeat.pid
+      beat -l info --pidfile ${SCHEDULER_PIDFILE}
     ;;
   monitor)
     [ -f ${MONITOR_PIDFILE} ] && rm ${MONITOR_PIDFILE}

[jezdez]

And also:

@@ -321,6 +323,16 @@ LOGGING = {
             'handlers': ['console'],
             'propagate': False,
         },
+        'celery.task': {
+            'level': 'DEBUG',
+            'handlers': ['console'],
+            'propagate': False,
+        },
+        'redbeat.schedulers': {
+            'level': 'DEBUG',
+            'handlers': ['console'],
+            'propagate': False,
+        },
         'request.summary': {
             'level': 'DEBUG',
             'handlers': ['console'],

[jezdez]

But only if you find those useful :)

[wlach]

These do indeed look useful! I'll add them.