Show us your Network Atlas maps!

[grantallenby]

Post of the week for me here

[shuth]

Basically, everything zackm wrote! And I've emailed you some screenshots.

• It is painfully slow to use!
• Should have capability to enable interface bandwidth/speed by assigning an interface to a connection (rather than relying on network topology and Connect Now - if SolarWinds doesn't know the topology, you don't get link information)
• Add the ability to add various objects to the page that don't have statuses - i.e. if you want to add a large rectangle to section off a part of the network, you have to create the background image and use it as the image background. This makes it frustrating to edit.
• Send to front/back, etc for layering
• Have I mentioned how painfully slow it is?
• It crashes more than half the time or locks up and stops responding. I've even got one customer where you can create a wireless heatmap, but you can't edit it once the heatmap is generated. If you try to open the map in NA, NA locks up and even after leaving it for an hour, you give up and have to force close. If you need to make a change, too bad - you have to delete the old one and start a new map and hope you don't make a mistake.
• A snap-to grid would be nice but I typically end up using the align tools

[ecklerwr1]

Wow I haven't had the map that wouldn't open again problem... that's bad.

[jblankjblank]

Thank you everyone for the feedback.  zackm I love the detail.  I wanted to check with you on point 7 you brought up: 

1. UnDP objects do not show the status per threshold/value, but the status of the ability of SW to poll an OID. (e.g.; if I have a UnDP that measures battery status on a UPS, the NA icon would be green as long as the SNMP query worked, no matter what the actual returned value was/is)

Have you tried the following?  View UnDP status on maps and Define UnDP Warning and Critical thresholds

This should allow you to specify thresholds for the UnDPs which will reflect status in an Atlas Map.  In summary:  Settings > Thresholds & Polling > Custom Poller Thresholds -

[zackm]

That's some good stuff!

Thanks for sharing, I would mention though that this is still limited to thresholds per a poller type, and not per poller.

e.g.;

Threshold for RemoteTemp NOT threshold for RemoteTemp on NodeA and a different threshold for RemoteTemp on NodeB

So, while this definitely addresses item 7 in my list; it also spawns a replacement.

[jblankjblank]

Absolutely - I think we always have an opportunity to do better!  We are certainly always trying to improve.  I also had another thought around this.  If the end goal is to get the Higher Level object (lets say a node) to change status, you could perform this with an alert.

Here is a general example (I used a quick UnDP I had in my lab already):

Within the alert you would need to take advantage of the Set Custom Status action:

Now, this currently only works for nodes, but you can get fancy with your Trigger Logic to get the desired result:

This will mark the node in a critical state which should draw your attention in the "map" or visual breakout.

Just a thought and something that is working for me in my lab. 

Thanks again for the feedback and insight as always!  

[mesverrum]

I do like this work around since critical is not normally used for node statuses it wouldn't be as confused with another condition.  Still kind of a clunky process until the atlas application itself supports grabbing the value you need.  In any case I think it's super important that custom sql/swql queries be supported in the same way as they are in email messages that way if people find gaps in the products in the future they can remediate it themselves rather than waiting potentially years for it to get added to the dev agenda.

[designerfx]

jblankjblank​ I tried that with a UNDP based alert and it simply triggered the alert on every single node in the node criteria, even after reproducing as above.   Are you sure your lab is/isn't beta?

[jblankjblank]

designerfx Positive this lab is not Beta - Unfortunately... you seem to be also correct around the behavior here - This is likely due to the separation of UnDP data in the database.  So, we are going to have to do better.  For now, you would have to create individual alerts for each device.  I also attempted to set the following:

Which did not appear to assist.  It sounds like the separate alerts would have to be done. 

Obviously not ideal, but try this and let me know if this works.

[designerfx]

Yeah we have 200 nodes which would be impacted I could probably SWIS creation of 200 alerts based on node URI, but I don't know if I'd consider that worth the effort involved when we barely have 1 trigger/week on average.