LTM iControlREST API request fails with HTTP 502 Bad Gateway

I'm running a viprion with multiple guests, all on 11.6.0 HF4. We see the same problem, where REST starts returning the 502 error OR "No JSON object could be decoded".

In our case, it seems the restjavad* logs will have a lck but no update for a period of time. Doing a grep on restjava, there will be a process for /usr/lib/jvm/java-1.7.0-openjdk-1.7.0.19.x86_64/jre/bin/java (with args), with a process start time equal to or near the log last entries. I've found that HUP'ping that process allows REST API to again be successful.

I've only done this on the standby unit - not the active. I do not know if this would negatively impact any operations, as I don't know what other TMM processes or OS services have a dependency. BUT.. maybe this will trigger a thought in some really smart people..

If I have time I'll scrape the logs for the same period as the HUP to see if anything reports a failure. My fear is that the customer/client connections would suffer silently.