From ced6c527bac72efe7537b846a6cec9c3bca8f6f4 Mon Sep 17 00:00:00 2001
From: msukkari <michael.sukkarieh@mail.mcgill.ca>
Date: Fri, 21 Feb 2025 17:13:35 -0800
Subject: [PATCH] add cpu split logic and only wait for postgres if we're going
 to connec to it

---
 Dockerfile       |  5 ++++-
 entrypoint.sh    | 51 ++++++++++++++++++++++++++++++++++++------------
 supervisord.conf |  5 ++---
 3 files changed, 45 insertions(+), 16 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index ff5eff9d..707bd57a 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -87,6 +87,9 @@ ENV DATABASE_URL="postgresql://postgres@localhost:5432/sourcebot"
 ENV REDIS_URL="redis://localhost:6379"
 ENV SRC_TENANT_ENFORCEMENT_MODE=strict
 
+# @nocheckin: we likely want to have a better solution for the self host case
+ENV BACKEND_CORES_FACTOR="0.5" 
+
 ARG SOURCEBOT_VERSION=unknown
 ENV SOURCEBOT_VERSION=$SOURCEBOT_VERSION
 RUN echo "Sourcebot Version: $SOURCEBOT_VERSION"
@@ -112,7 +115,7 @@ ENV POSTHOG_PAPIK=$POSTHOG_PAPIK
 ENV STRIPE_PUBLISHABLE_KEY=""
 
 # Configure dependencies
-RUN apk add --no-cache git ca-certificates bind-tools tini jansson wget supervisor uuidgen curl perl jq redis postgresql postgresql-contrib openssl
+RUN apk add --no-cache git ca-certificates bind-tools tini jansson wget supervisor uuidgen curl perl jq redis postgresql postgresql-contrib openssl util-linux 
 
 # Configure zoekt
 COPY vendor/zoekt/install-ctags-alpine.sh .
diff --git a/entrypoint.sh b/entrypoint.sh
index aa01ffb1..c281843a 100644
--- a/entrypoint.sh
+++ b/entrypoint.sh
@@ -181,25 +181,52 @@ echo "{\"version\": \"$SOURCEBOT_VERSION\", \"install_id\": \"$SOURCEBOT_INSTALL
 
 
 # Start the database and wait for it to be ready before starting any other service
-su postgres -c "postgres -D $DB_DATA_DIR" &
-until pg_isready -h localhost -p 5432 -U postgres; do
-    echo -e "\e[34m[Info] Waiting for the database to be ready...\e[0m"
-    sleep 1
-done
+if [ "$DATABASE_URL" = "postgresql://postgres@localhost:5432/sourcebot" ]; then
+    su postgres -c "postgres -D $DB_DATA_DIR" &
+    until pg_isready -h localhost -p 5432 -U postgres; do
+        echo -e "\e[34m[Info] Waiting for the database to be ready...\e[0m"
+        sleep 1
+    done
 
-# Check if the database already exists, and create it if it dne
-EXISTING_DB=$(psql -U postgres -tAc "SELECT 1 FROM pg_database WHERE datname = '$DB_NAME'")
+    # Check if the database already exists, and create it if it dne
+    EXISTING_DB=$(psql -U postgres -tAc "SELECT 1 FROM pg_database WHERE datname = '$DB_NAME'")
 
-if [ "$EXISTING_DB" = "1" ]; then
-  echo "Database '$DB_NAME' already exists; skipping creation."
-else
-  echo "Creating database '$DB_NAME'..."
-  psql -U postgres -c "CREATE DATABASE \"$DB_NAME\""
+    if [ "$EXISTING_DB" = "1" ]; then
+        echo "Database '$DB_NAME' already exists; skipping creation."
+    else
+        echo "Creating database '$DB_NAME'..."
+        psql -U postgres -c "CREATE DATABASE \"$DB_NAME\""
+    fi
 fi
 
 # Run a Database migration
 echo -e "\e[34m[Info] Running database migration...\e[0m"
 yarn workspace @sourcebot/db prisma:migrate:prod
 
+# Get total CPU cores
+TOTAL_CORES=$(nproc)
+echo -e "\e[34m[Info] Total CPU cores available: $TOTAL_CORES\e[0m"
+
+# Default to 50/50 split if BACKEND_CORES_FACTOR not set
+BACKEND_CORES_FACTOR=${BACKEND_CORES_FACTOR:-0.5}
+
+# Calculate number of cores for backend (rounded down)
+BACKEND_CORES=$(printf "%.0f" $(echo "$TOTAL_CORES * $BACKEND_CORES_FACTOR" | bc))
+
+# Ensure at least 1 core for each
+if [ $BACKEND_CORES -lt 1 ]; then
+    BACKEND_CORES=1
+fi
+if [ $BACKEND_CORES -ge $TOTAL_CORES ]; then
+    BACKEND_CORES=$(($TOTAL_CORES - 1))
+fi
+
+# Generate CPU lists
+export BACKEND_CPU_LIST=$(seq -s, 0 $(($BACKEND_CORES - 1)))
+export FRONTEND_CPU_LIST=$(seq -s, $BACKEND_CORES $(($TOTAL_CORES - 1)))
+
+echo -e "\e[34m[Info] Backend will use CPUs: $BACKEND_CPU_LIST\e[0m"
+echo -e "\e[34m[Info] Frontend will use CPUs: $FRONTEND_CPU_LIST\e[0m"
+
 # Run supervisord
 exec supervisord -c /etc/supervisor/conf.d/supervisord.conf
\ No newline at end of file
diff --git a/supervisord.conf b/supervisord.conf
index ea746564..b2dcc53a 100644
--- a/supervisord.conf
+++ b/supervisord.conf
@@ -4,7 +4,7 @@ logfile=/dev/null
 logfile_maxbytes=0
 
 [program:zoekt]
-command=./prefix-output.sh zoekt-webserver -index %(ENV_DATA_CACHE_DIR)s/index -rpc
+command=taskset -c %(ENV_BACKEND_CPU_LIST)s ./prefix-output.sh zoekt-webserver -index %(ENV_DATA_CACHE_DIR)s/index -rpc
 autostart=true
 autorestart=true
 startretries=3
@@ -13,7 +13,7 @@ stdout_logfile_maxbytes=0
 redirect_stderr=true
 
 [program:web]
-command=./prefix-output.sh node packages/web/server.js
+command=taskset -c %(ENV_FRONTEND_CPU_LIST)s ./prefix-output.sh node packages/web/server.js
 autostart=true
 autorestart=true
 startretries=3
@@ -21,7 +21,6 @@ stdout_logfile=/dev/fd/1
 stdout_logfile_maxbytes=0
 redirect_stderr=true
 
-
 [program:backend]
 command=./prefix-output.sh node packages/backend/dist/index.js --cacheDir %(ENV_DATA_CACHE_DIR)s
 autostart=true