Production and Staging have been experiencing downtime

[gbinal]

In recent months (it seems like 3-4 times a month) either production or staging will go down (site returns a 500 error with a message saying (e.g.): 404 Not Found: Requested route ('pulse.app.cloud.gov') does not exist.).

The New Relic alerts are catching it and a simple restart of the app on cloud.gov gets it back up, but obviously, this is not a good thing. Unfortunately, in the last week, it's been happening more often.

We're investigating the causes, but some initial ideas for solving it include:

• bumping up the memory
• add cache-breaking (e.g. after ?) after deploy; to help test the effects of deploy

[micahsaul]

Looking at New Relic, it seems like all of the errors I'm seeing are related to someone probing us for vulns. The question, though, is why that would cause the whole site to crash.

[konklone]

I would recommend we bump up the memory.