Week 09

Troubleshooting 01.04.24

It's been the second time the droplet logged us out and the graphs provided for overall monitoring from the teacher show that the latest output was not tracked by the diagram.

• issue 1: latest file accessed concurrently. Solution was to add latest to the db.
• issue 2: we have observed that the disk space raised by 15 % the past 12 hours. I traced it to the docker container, which is surprisingly our minitwit container (and not prometheus). Having the logging level set to Trace for both "Default" and "Microsoft.AspNetCore" in our .NET Core application's appsettings.json is the most verbose logging level and will generate a very high volume of log data. Trace logging includes detailed information about the application's behavior and is typically only recommended for debugging purposes in a development environment. In a production environment, this level of verbosity can quickly consume disk space and might even impact performance. Therefore logging level is set up to "warning"
• issue 3: We choose to host the Grafana and Loki on a separate node in our system, such that it would be fault-tolerant, and give us the information needed in case of a break down on our application. Digital-ocean provided an implementation through their marketplace, which we implemented. However, this gave us some different complications, which we later learned was due to it being an old version. In an effort to update it, we found that the implementation was very integrated in some api from digital ocean. We found it hard to configure it, and decided to set the server up on a new droplet ourselves in order to keep the control. We have added a volume to store historical data. Grafana is now up and running on the new droplet, pointing to the production server. The original reserved-ip has been assigned to the new self-created server. The Dashboard could still need some queuring and configuration to give the best monitoring of our application.