[Fixed]-Sudden 503s on OpenShift Django. Need help debugging

I know you mentioned that there’s nothing in the logs, but I would anyway try tailing all the logs rhc tail <yourApp> and watching in real time for any clues there, when the 503’s are returned.

To check whether your gear is not restarting due to insufficient memory, I recommend this.
Having your ssh connection closed unexpectedly may be another indicator of unexpected gear restarts.

Note that htop displays only your tasks, which take only little resources in context of the whole node; using e.g. 3% of memory of 16GB may be nearing the small gear’s limits (512 MB).