fix: increase flask backend memory limit to prevent OOM kills

The monitoring_flask_backend (flask-pgss-api) workers were being
SIGKILL'd by the OOM killer every ~5 minutes — exactly when
VictoriaMetrics scraped /query_info_metrics. With 4 gunicorn workers
and 4750+ pgss query rows to process, 192MB was insufficient.

Bump mem_limit from 192m to 384m.

Relates to https://gitlab.com/postgres-ai/postgresai/-/issues/148

https://claude.ai/code/session_01SzJxzZNQjDQphaHyaX3RU7

Author: claude

docker-compose.yml

@@ -200,7 +200,7 @@ services:
     image: ${PGAI_REGISTRY:-postgresai}/monitoring-flask-backend:${PGAI_TAG:?PGAI_TAG is required}
     container_name: flask-pgss-api
     cpus: 0.1
-    mem_limit: 192m
+    mem_limit: 384m
     environment:
       - FLASK_ENV=production
       - PROMETHEUS_URL=http://sink-prometheus:9090