Add notification status "past"

[Labsy]

Hi,

 

I have a lot of ping and cpu notifications, but I must read one by one to see, which one is still critical, and which is somehow resolved. So I suggest to have additional notification status, like "past" or something, which would be set for notification, where there is no alarm state anymore and the notification has not been read nor deleted.

Server/PC status would not have red or yellow dot anymore for such statuses (meaning, at the moment I do not have to worry about server/PC)

 

Thoughts?

[Paul]

Hi,

So you're asking that instead of a new notification that the ping endpoint can be reached again you would like that the agent would modify the previously sent ping failure notification to change it's priority and content.

I like the idea but I don't see a practical use of it. When I receive a notification that some device cannot be reached I would like to know why that happened and queue an investigation of the network to evaluate potential downtimes and future risks. At the moment notifications can be ordered by date so I can quickly identify when something happened and when it got fixed (if applicable).

However this doesn't mean that the notification cannot change a notification if the problem got fixed as long as the two notifications get merged nicely (Fix date to be properly appended to the original message and notification priority to be dropped to low or normal).

Edited October 13, 2013 by Paul

[Labsy]

Hi Paul,

 

what you say is very true, so actually I reconsidered my request. But I do not have any good idea how to solve.

What bothers me is a dozen of notifications on my android phone, most of them about PING and some about CPU, elevated. For example, I have 5 notifications about PING elevated for past 4 hours, but when I look at server there's nothing wrong. Even more - if I look at MRTG on that server, I cannot see any traffic congestions, no ping delays, which is weird.

 

So I assume either PCM hicups and does not measure PING properly, or the issue was transient and poses no threat to server functionality.

...or at least but not last, I might be wrong and I have some hidden problems on server  

[Paul]

What is the ping threshold you configured in PCM Manager?

[Labsy]

Hmmm.... for example, if an average ping during busy hours is between 1 and 2 ms, then I set threshold to 5 ms in 5 minutes period. 

And that setting triggers a notification few times a day.

 

Might it be due to ping collection method?

For example, PCM may collect ping response in few different ways:

- might issue system PING command and only read response from system

- might have own PING module and ping from PCM directly

- also there is a posibility that PCM only sends 1 ping in time intervals, which should be corrected to at least 2 or 3 pings each time

[Paul]

5 ms is pretty bad I might say. On a loaded network 5 ms can happen whenever there is a considerable amount of traffic in that moment. For instance you might have a really nice 10/100/1000 network over there, all patched up fancy but spikes can occur due to switches and other transport issues.

 

5 ms is very sensitive...

 

From my research I found out that there already exists a Ping class inside .NET which offers the same functionality of the win32 ping application. Further more I dug up that the .NET ping class uses an unmanaged method IcmpSendEcho2 from the iphlpapi.dll which is the exact same method used by the native ping application judging from unmanaged method imports from a debugger's point of view. So it's fair to say that PC Monitor does that exact same thing as you would do when you test out your network connectivity so we can exclude the first and second possibilities you pointed out. For the third way there is no way of knowing unless someone from PC Monitor will say how their ping logic works.

Edited October 14, 2013 by Paul

[Labsy]

Well, back to the topic.

 

This morning I woke up and found notifications for 3 servers to be offline! Even before I opened my eyes completely, I held internal panic inside and started investigation, even though my view was still blurry.

But nothing... I found absolutely nothing wrong, all servers were up and running, no complaints from users, no other alarms.

 

Hmmm....must be a glitch either in PCM agent, maybe some hiccup on internet connection....

...but the fact at the moment when I looked at the screen of my mobile phone was, that at that particular moment servers were NOT OFFLINE ANYMORE.

 

Well, for such cases I still think some "auto-update" feature might be built-in to PCM to automatically mark such a notifications as pase, maybe some other color, not red anymore. With other words, the color of notification might change depending on CURRENT item status - while elevated or critical state lasts, the appropriate color should be applied to notification. But when this state is over, also notification color should change.

 

Too complicated?

What do you think, guyz?

[Paul]

AutoUpdate will stop the services starting the offline countdown on the cloud servers but you should only be getting an offline notification if the agent is unreachable for something like 10-15 minutes I think.

[Labsy]

Hmmm....yes, that's true.

But still my logic is to have server with RED color only for the time being unreachable. When it comes back online, I expect RED bubble should disappear from this server in main screen.

[Paul]

True that, however I still need to know that something bad happened at a moment thus the red light still being there. I need to know something went wrong and it goes back to normal only after I remove the notification meaning that I have solved the issue or have understood what the problem was.

 

I guess an option to clear notification if another positive one occurs would be nice, or maybe to merge the positive one with the negative and lowering it's priority.

[geislinn]

Hi. Has this been implemented yet??

[Paul]

No, the notification system did not change. If you want, you can send an email at: support at mobilepcmonitor dot com, requesting this Auto Clearing Notification feature. More people ask for it, more importance this feature will receive.

Edited March 15, 2014 by Paul