Currently, we're using Graphene-Django API on an Azure Virtual Machine that's running Ubuntu 18.04 LTS and Maria DB. We also have a React client that runs in a WordPress environment that interacts with this API. The client uses the Apollo boost package, and currently we're only using queries without any mutations. We can normally handle the traffic, and it's been running fine for about 6 months now. However, whenever traffic surges the API really starts to chug, and slows down drastically, to the point where queries can take upwards of a minute to complete. When the server is rebooted, the problem goes away.
Has anyone ever experienced anything like this before? I've researched this extensively, and can't seem to find any answers.