Release: Bug Battle
Welcome to The Silly Home’s 5th release - Bug battle
This is a long overdue release but I had been busy travelling! This is a summation of a few changes I made based on the support I was giving in our discord channel.
Don’t feel shy to post any comments, criticism and ideas below. I want the feedback and ideas to make this amazing. We also have a discord chat if anyone needs help setting up or just wants a chat.
One of our users indicated that the container was using an outrageous 15gbs of RAM on start-up causing the container to shutdown due to lack of resources. I did some investigation with the results noted in this post. The changes to the query helped reduce the load.
To assist in user config, I’ve added a few checks so that things like spelling mistakes can be found quicker in the execution.
Better logging for addon
It is quite hard to get logs from The Silly Home Addon. The container itself is unreachable on HASS so the local log file cannot be accessed meaning you can only see it in the UI. However, the logs themselves are often flooded by AppDaemon messages, so the underlying cause is hard to find. I have added 2 things:
- Sending logs and config to me. With adding a username and password field to the config, I can more easily identify the logs of particular users which can help me debug. The usernames and password (>8 chars) needs to be unique. All security info is not sent!
- Don’t start AppDaemon if the model generation failed.
Removal of sharing data by SQL
The sharing of the states data is not very efficient with SQL and is causing the RAM issues. In the future when moving to AWS, a flat file will suffice through API gateway. For now, I’ve removed the states data sharing through SQL.
I think for now, having a production and dev version isn’t adding too much value (as there are too many updates and fixes). As of now, I intend to only update the DEV release until further notice.
Next week - Production and Beyond
I will follow up back onto the productionalization of The Silly Home container! The AWS Sagemaker piece is almost complete. Now to create API gateways to service the model training.