fix: add 30min stop_grace_period to prevent data corruption on shutdown

FreqAI training can take minutes to hours. Docker's default 10s grace
period causes SIGKILL mid-write, corrupting feather/pickle files.
Give freqtrade up to 30 minutes to finish training and flush data
before Docker sends SIGKILL.

diff --git a/ReforceXY/docker-compose.yml b/ReforceXY/docker-compose.yml
index 2c956e54a0299339c0d8621ac02c7b58d5278e1a..98df218115d37be70ed8131cba80042b496598bd 100644 (file)
--- a/ReforceXY/docker-compose.yml
+++ b/ReforceXY/docker-compose.yml
@@ -18,6 +18,7 @@ services:
       dockerfile: Dockerfile
       args:
         optuna_version: 4.8.0
+    stop_grace_period: 1800s
     restart: unless-stopped
     container_name: freqtrade-ReforceXY
     environment:

diff --git a/quickadapter/docker-compose.yml b/quickadapter/docker-compose.yml
index 2b5d005bc39997293b172ee5735be6a519a8a62c..496982449e3a16d7f049d37bf79f5ce5f93a605d 100644 (file)
--- a/quickadapter/docker-compose.yml
+++ b/quickadapter/docker-compose.yml
@@ -22,6 +22,7 @@ services:
         scikit_image_version: 0.26.0
         ngboost_version: 0.5.10
         catboost_version: 1.2.10
+    stop_grace_period: 1800s
     restart: unless-stopped
     container_name: freqtrade-quickadapter
     environment: