I've just started taking a look at the open source version, so hopefully I'm not missing the mark too much on some of this. I like the product overall, but some adjustments would be welcome.
1. When you change monitoring defaults, it appears to apply the changes to existing monitored resources. That should be an option when you update the defaults. I would hate to tweak 50 resources to my liking, then need to make a minor change to the defaults, wiping out my tweaks.
2. I don't see any SNMP auto-discovery options. Maybe I missed it. It should be able to scan ip ranges periodically, importing devices with the configured SNMP string(s).
3. Where do I configure the default SNMP string for new platforms (specifically Cisco IOS)? The only way I've see to do this is to change it on each platform after creation. Also, the default SNMP agent address is the loopback address, which will fail every time. It should default to the platform's designated ip address.
4. It should be possible to re-scan an SNMP platform for new interfaces, etc. If I add a new blade to a switch, I don't see any way for it to pick up those new interfaces. The Auto-Discovery function only appears to apply to the agent based platforms.
5. In the resource list, why display the cryptic name for resources if I've given it a descriptive name in inventory? When you're looking at a platform with hundreds of resources, it's tough to wade through the list. I might want to name an interface 'New York - MetroE Interface' or something easier to find. Currently, I have to look through hundreds of names like: "10.11.1.1 ny-mdf-4506 GigabitEthernet2/5 Interface (0:b:5f:bc:0:f2)". That's cool for initial discovery, but not for day to day use.
6. Related to #4, the search function should search all descriptions as well as their actual names.
7. It should be possible to tag a resource in inventory as 'Do Not Display/Monitor' or something like that. If you have a Cisco switch with 500 interfaces, you generally only care to look at a select few. I accomplished something like this by deleting the unwanted interfaces from the platform's inventory, but that doesn't seem to be a great way to go about it. If you decide later to monitor one that was deleted, it seems a little tough to add it back. You should be able to go to inventory and check a box to for 'Hide from Resource List/No Monitoring'.
8. The default monitoring template should be separate from the notion of whether a resource is being monitored. For example, by default I don't want any Cisco IOS interface to be monitored, since you're talking about tons of interfaces. But when I do want to monitor a particular interface, I would like to enable it and it inherit the default settings from the template. As it is, I have to disable all monitoring in the template, and then on specific interfaces, enable the metrics remembering to set them the same as other monitored interfaces.
9. When editing a monitoring template, the 'Default On' setting needs to be independent from the 'Collection Interval' settings. As it is, at least from what I can tell, to enable collection of a metric, there's no 'Enable Collection' button to match the 'Disable Collection' button. You enable collection by setting the 'Collection Interval'. That's not too bad if all of your intervals are the same, but what if you have a different interval for each metric? As simple 'Enable Collection' button would be handy.
10. Saved charts need a descriptive name just like resources. I like the 'Save Chart to Dashboard' feature, but the display name is painful. Instead of '10.11.1.1 ny-mdf-4506 GigabitEthernet2/5 Interface (0:b:5f:bc:0:f2): Multiple Metric', I would like to call it 'NY Internet Feed Throughput' or something like that.
11. I may be misinterpreting this chart data, but when I look at something like 'Bits Received per Second' on a Cisco IOS interface, it appears that the chart's scale is in Bytes rather than Bits. The low, average and peak stats are noted in Mbps, which is good, but it's not nice to always have to break out the calculator to convert Bytes back to Bits to see what the chart is telling you. The chart should match the units of the metric. Show me Mbps, Kbps, or whatever - just not bytes.
12. On charts, it would be cool to be able to hover over a data point and see the actual measurement. You can sort of gauge it approximately. Just a nice addition.
I've focused more on Cisco IOS than anything so far, and it seems that the product is better suited to monitoring servers and applications, but most of these comments carry over to other platforms. Of course it's not likely that a Windows server would have 500 network interfaces, so the resource view isn't quite as cluttered, but it would still be nice to label things more clearly and hide things from the foreground, like the Windows loopback adapter. There are various other issues that I left out because I think they're part of the subscription product. If I were going to subscribe to the Enterprise version, this kind of stuff would have to be in there.
Having said all that, I like the basic functionality of the product. It's got great potential. Sorry if I missed the mark on any of this.
This is great feedback. You are not off the mark at all. In fact, all of your points are valid enhancement requests. We need feedback like this to better suit HQ to the people who will be using it. I am going to start logging these in our issue tracking system. I will post the links to the issues when I am done.
The ability to give a descriptive display name I wish for, as well.
#11, Bps vs. bps, I see the same inconsistency.
#12, EMC's Navisphere has this, and it's quite nice
#3, not only does SNMP default to local loopback, but so do the default JDBC connect strings for Sybase, Oracle, etc., and the td:// connect strings for WebLogic. There may be others, but I am specifically aware of these. Additionally, if you can detect all the Sybase/WebLogic/what-have-you install stuff, the JDBC string should automagically default to the appropriate port #.
The numbers in the list below match up to your list of enhancements. I logged new enhancement requests for most. #5 was already logged, I just made a note in it to be sure to cover your request. There are a few that I did not log:
6. I was able to search on the description and it worked. Does this not work for you? If you just mean that you should be able to search on the real name and any custom name you give to a resource, I can make a note of that in the logged issue.
7. The solution here could be a combination of the solution in #8 used in conjuntion with the role based authorization feature already available in the enterprise edition of HQ. Essentially, you would group the resources you care about, apply the metric template you want them to have (see the link for #8 below) and then put that group of resources in the Role assigned to the user you will be using to log into HQ. You could then, optionally group the resources you don't care about, apply another metric template to that group which would basically be empty so there would be no metrics collected for those resources. This would make it so you see only the resources you care about and the ones you do not care about are hidden and not being monitored. Yet they are still in the inventory, so if you needed to at any time, you could enable them and make them visible.
11. This sounds more like a bug than an enhancement request. I may need to get more details on this, but it sounds like the chart is named "Bits..." yet it is displaying data in bytes. Is my understanding correct?
Quoth John: 11. This sounds more like a bug than an enhancement request. I may need to get more details on this, but it sounds like the chart is named "Bits..." yet it is displaying data in bytes. Is my understanding correct?
This correctly describes the behavior that I am observing. Metric is bps, chart is displaying Bps.
Sorry, I've been sidetracked. My installation mysteriously stopped working, so I'll need to check into that before I can answer about #6. If it works for you it may have been a quirk on my trial.
If that solution for #7 is part of the subscription version, that's cool. It makes things a little tough in my case. Even if I can't filter down that list, if it could display the friendly name that I assign, I could quickly find the interfaces that I'm interested in.
Your understanding of #11 is correct. If you chart a Cisco IOS interface with any 'bits per second' metric, the chart will be displayed in bytes.