ci: Check health of services after running integration tests and fix snuba-replacer

Also remove batching from snuba-replacer. This was removed in getsentry/snuba@6e98074

Co-authored-by: Chad Whitacre <[email protected]>

Author: emmatyping

_integration-test/run.sh

@@ -135,3 +135,18 @@ source ./custom-ca-roots/setup.sh
 $dcr --no-deps web python3 /etc/sentry/test-custom-ca-roots.py
 source ./custom-ca-roots/teardown.sh
 echo "${_endgroup}"
+
+# Table formatting based on https://stackoverflow.com/a/39144364
+COMPOSE_PS_OUTPUT=$(docker compose ps --format json | jq -r \
+  '.[] |
+   # we only care about running services. geoipupdate always exits, so we ignore it
+   select(.State != "running" and .Service != "geoipupdate") |
+   # Filter to only show the service name and state
+   with_entries(select(.key | in({"Service":1, "State":1})))
+ ')
+
+if [[ "$COMPOSE_PS_OUTPUT" ]]; then
+  echo "Services failed, oh no!"
+  echo "$COMPOSE_PS_OUTPUT" | jq -rs '["Service","State"], ["-------","-----"], (.[]|[.Service, .State]) | @tsv'
+  exit 1
+fi

docker-compose.yml

@@ -260,7 +260,7 @@ services:
     command: consumer --storage transactions --consumer-group transactions_group --auto-offset-reset=latest --max-batch-time-ms 750
   snuba-replacer:
     <<: *snuba_defaults
-    command: replacer --storage errors --auto-offset-reset=latest --max-batch-size 3
+    command: replacer --storage errors --auto-offset-reset=latest
   snuba-subscription-consumer-events:
     <<: *snuba_defaults
     command: subscriptions-scheduler-executor --dataset events --entity events --auto-offset-reset=latest --no-strict-offset-reset --consumer-group=snuba-events-subscriptions-consumers --followed-consumer-group=snuba-consumers --delay-seconds=60 --schedule-ttl=60 --stale-threshold-seconds=900