Pi3 weird behavior - sleep-like symptoms

[ryebrye]

I have a pi 3 set up in a system where I have a java-based app running and monitoring a couple of switches and activating a few relays. More specifically it's being used to control a vacuum system for maple sugaring, and it monitors float switches in a chamber and activates a three-way valve to let air in (and let sap out) when the chamber fills.

This has been working fine for quite some time - a few little glitches here and there, but nothing major. It was running on a pi2 for most of the time, I just switched it over to a pi3.

In the past I've had issues where the system didn't quite work because of loose connection etc - so it didn't get the signal for the sap filling the chamber up and it would fill up a moisture trap I have up the line from it.

Today something even stranger happened. I was following the twitter feed of it and noticed that it was normally dumping sap every 10 minutes, but something happened and it had been about 30 minutes since the last time it dumped sap. This was odd - so I ssh'ed into my home router and then ssh'ed into the pi to check the logs.

When I checked the logs, I noticed the following:

2016-03-09 10:53:30.197 DEBUG 466 --- [ent-executor-11] com.ryebrye.releaser.HardwareImpl : low float switch signal is now false
2016-03-09 10:53:30.200 ERROR 466 --- [itchStateChange] com.ryebrye.releaser.ReleaserControl : Sap below low point is not a valid trigger for the state of Filling
2016-03-09 10:53:59.768 DEBUG 466 --- [ent-executor-11] com.ryebrye.releaser.HardwareImpl : low float switch signal is now true
2016-03-09 10:53:59.774 ERROR 466 --- [itchStateChange] com.ryebrye.releaser.ReleaserControl : sap above the low point, but that's not valid for the state of Filling
2016-03-09 11:25:03.492 DEBUG 466 --- [ent-executor-12] com.ryebrye.releaser.HardwareImpl : high float switch signal is now true
2016-03-09 11:25:03.498 INFO 466 --- [itchStateChange] c.github.oxo42.stateless4j.StateMachine : Firing SapAboveHighPoint
2016-03-09 11:25:03.500 INFO 466 --- [itchStateChange] com.ryebrye.releaser.ReleaserControl : sending open message

It's important to note that I logged in at 11:25 AM and then almost _immediately_ after that it detected the high switch being active (the event listener for the GPIO state change happened) and then went into it's dumping routine and it read the low float switch states as the sap drained out etc...

It's seems highly circumstantial that my SSH'ing into the pi would line up exactly with the time that it read the GPIO states and decided to dump out the sap.

Is there some kind of sleep mode or some kind of low power state it goes into? Is there any way to disable it?

In addition to switching to the pi 3 I also updated to Raspbian Jessie, so it's possible that it's not a pi-3 specific issue but might be a jessie issue.