Configuration issues [fault] announcement background deploy new blog at risk of failure

Recently, our new blog background ( Angular 8.2.7 + the .NET 3.0 Core conduct) issued gray, jumps to the background if you visit the blog https://i-beta.cnblogs.com/ , instructions for use is The new blog background.

Today we are in a gitlab-ci-based automated publishing process, due to operational problems no modification appsettings.Production.json be saved before release, causing the container to use an older version of the configuration file at startup, plus the container's health check can not check out this abnormal situation (improved this place is not yet complete), most should be tested to verify that there is no duty to and after the release of the key functions is not timely processing of user feedback, resulting in 18:22 to 19 : after the user to use the new blog can not be published during the 27 normal Bowen, very sorry this may have caused trouble, please understand.

We will learn, and take the following improvements:

  • Higher priority to improve the health check. One vessel health check, and so second is Ali monitoring of health checks. When the key function is not available, so that health check fails (no previous health check on the business function check). If a problem occurs when this release, container health check fails, docker swarm will not deploy new container. When running the container problems affect the use of the key functions of the police in time.
  • Achieve the highest possible verification testing of key functions with "robot" in a production environment after release.
  • Every automated message to inform the duty officer on duty bulk note when publishing user feedback.

Guess you like

Origin www.cnblogs.com/cmt/p/11863251.html