Ensure Critical Errors are logged Before Termination

[pleshakov]

Currently, when the code panics, we get the following in the logs:

panic: unknown file type 1
goroutine 220 [running]:
github.com/nginxinc/nginx-kubernetes-gateway/internal/mode/static/nginx/file.ensureType(...)
        github.com/nginxinc/nginx-kubernetes-gateway/internal/mode/static/nginx/file/manager.go:148
github.com/nginxinc/nginx-kubernetes-gateway/internal/mode/static/nginx/file.writeFile({0x1bba038?, 0x27cac98?}, {{0xc00049e080, 0x3a}, {0xc0003b8800, 0xa8e, 0xa8e}, 0x1})
        github.com/nginxinc/nginx-kubernetes-gateway/internal/mode/static/nginx/file/manager.go:116 +0x4d9
github.com/nginxinc/nginx-kubernetes-gateway/internal/mode/static/nginx/file.(*ManagerImpl).ReplaceFiles(0xc0002d01c0, {0xc0004c33e0, 0x2, 0x1?})
        github.com/nginxinc/nginx-kubernetes-gateway/internal/mode/static/nginx/file/manager.go:104 +0x32d
github.com/nginxinc/nginx-kubernetes-gateway/internal/mode/static.(*eventHandlerImpl).updateNginx(0xc000174800, {0x1bb9348, 0xc00026b900}, {0xc00048b6e0, {0xc00046e580, 0x1, 0x1}, {0xc0004a4000, 0x2, 0x2}, ...})
        github.com/nginxinc/nginx-kubernetes-gateway/internal/mode/static/handler.go:85 +0x9e
github.com/nginxinc/nginx-kubernetes-gateway/internal/mode/static.(*eventHandlerImpl).HandleEventBatch(0xc000174800, {0x1bb9348, 0xc00026b900}, {0xc00047e500?, 0x14, 0xc00012f720?})
        github.com/nginxinc/nginx-kubernetes-gateway/internal/mode/static/handler.go:71 +0x210
github.com/nginxinc/nginx-kubernetes-gateway/internal/framework/events.(*EventLoop).Start.func1.1({0xc00047e500, 0x14, 0x14})
        github.com/nginxinc/nginx-kubernetes-gateway/internal/framework/events/loop.go:68 +0x10a
created by github.com/nginxinc/nginx-kubernetes-gateway/internal/framework/events.(*EventLoop).Start.func1
        github.com/nginxinc/nginx-kubernetes-gateway/internal/framework/events/loop.go:65 +0x14a

Acceptance criteria:

• Log an error before exiting so that it can be parsed by the logging systems.

[mpstefan]

We determined that this case should be exceptionally rare, and even in the event this occurs, the log output right now should be sufficient to track down whatever problem occurred.