MX Firewall - Lost connectivity vs. lost power?

[cburtbcit]

Does anyone know of a way to determine if an MX appliance went offline due to losing power versus losing connectivity? The event log doesn't seem to contain a "boot" event, only primary uplink status change. Presumably, this event won't show up in the log if there's no secondary link to transmit it.

[Adam · Accepted solution]

One of the many reasons device uptime should be available in the UI to see for MX, switch, and APs.  You can call support to have them verify the uptime.  Also will be some info in Network Wide>Event Log. 

[cburtbcit]

Thanks for the reply. All I see in the event log is a gap in the events during the outage I'm investigating. I don't know if that gap exists because the device was unable to send log data to the cloud due to a connectivity problem, or if it's because the device wasn't powered on to log data.

I will follow your advice and contact support to get the up-time.

[PhilipDAth]

@Adam, while I was at Masters I also asked all the product teams to make their devices create an event log entry for when they are rebooted (and by whom).  Then at least you can tel the difference between a manually initiated reboot, and an unexpected outage.

[cburtbcit]

@PhilipDAth that certainly seems like an obvious feature to add. I found it amusing that I bragged to myself about how easy it would be to figure this out, just to find that the error log is chock full of information I may never use, and is missing this seemingly-critical piece of information.

I would add to your suggestion that at 15 or 30 minute intervals, there should be a general status log entry that includes the uptime, the number of errors since the last status entry, and perhaps avg. bandwidth utilization over the interval.

[Adam]

Uptime is available to support so I'm sure it wouldn't be very complex to add to the UI.  But an event log entry on bootup or even an alert option for any device that has booted up would be nice.  We had a failing switch that would randomly reboot during the day and it would come back up before it could even trigger the 5 minute offline alert so the only way we found out was from users reporting it.  

[cburtbcit]

In my case, I have a service provider telling me one thing, a user base telling me another, and equipment that isn't say much of anything other than there was an outage. So I think your suggestion would solve a lot of problems. I think it would be nice if the status graphs (the green bars with info on mouseover) could be updated to show that the device in question was rebooted as well, and wasn't simply disconnected. It's great to see a graph of availability, but it's just pretty colors if you have to comb through the event log for any real detail.