91 Replies Latest reply on Jul 9, 2012 7:10 AM by dclick

    If you're curious as to what we're working on...

    bshopp

      The NPM team is busy at work on a number of highly requested enhancements:

       

      • Scheduled Discovery - the ability to schedule discoveries on a recurring basis
      • Syslog and Traps –
        • Ability to reference Trap variable binding names and value pairs as variables in alert messages (e.g. ${vbName1}, ${vbData1})
        • Ability to configure Trap alerts based on pattern matching on Trap Details
        • Ability to reference Orion node variables in Syslog and Trap alerts
        • Respect Orion NPM unmanaged node status in Syslog and Trap alerts
        • Ability to change the status of an interface in Orion NPM based on Trap message
        • Add option when forwarding a Syslog message to retain the original source IP address
        • Add option to consume original address when passed in a Syslog message.
        • Add option when forwarding a Syslog message to retain (and spoof) the original source IP address.
      • Full VMWare vSphere Support - in order to add support for vSphere 4.0 from VMWare, we can no longer gather the data we need via SNMP so we are updating our integration to gather this via their API
      • Connect Now - the ability to drag one or more nodes on to a map and based on the last discovery topological data, automatically connect those device together on the map
      • Product update notification – an in product notice to alert you to new releases and service packs are available for your Orion products
      • Improved upgrade experience – additional effort to reduce upgrade disruptions
      • Official platform support for Windows 2008 R2
      • Network Atlas performance enhancements
      In parallel, we are also working on the following:

      FIPS support – ensures product operability on a Windows OS (2003 & 2008) with the FIPS Group Policy Object (GPO) enabled, which restricts which cryptographic algorithms are allowed

      PLEASE NOTE:  We are working on these items based on this priority order, but this is NOT a commitment that all of these enhancements will make the next release.  We are working on a number of other smaller features in parallel.   If you have comments or questions on any of these items (e.g. how would it work?) or would like to be included in a preview demo, please let us know!

        • Re: If you're curious as to what we're working on...
          justty

          This seems to be a popular request. Did this make it in?

          Blackout / Maintenance Windows

          • Re: If you're curious as to what we're working on...

            Not sure if this is addressed in the work being done for future release but will there be a way to more seemlessly integrate syslog and/or snmp traps into NPM?  For instance it would be nice to generate an NPN alert (or event) when NPM receives a specific syslog msg or specific trap.  The current version of NPM doesn;t seem to integrate syslog and/or snmp within NPM to this level.  I would like a snmp trap, for instance, to generate the same type of alert as wehn a managed interface goes down.  Will this be possible?

            • Re: If you're curious as to what we're working on...

              Brandon, all good stuff -- we look forward to the continued improvements.  The topology aware feature is very much needed!

               

              Thanks!

              • Re: If you're curious as to what we're working on...

                Brandon, something else to consider for features:

                When performing a discovery of a network we will run into network devices that are multi-homed (MH) and in cases of fairly large networks these multi-homed devices will have multiple Layer3 paths.  The issue we run into is that during the discovery each IP address, although discovered, ends up being presented as a separate device.  So for any one device I will end up with N instances, N being determined by the number of pingable interfaces.

                Granted, I can fix this by following best practices and enabling loopback then scanning the loopback network, but this only works for multi-homed devs that provide a loopback that is -- flexible (eg., Cisco).

                Ideally, I should be able to discover an MH device and during that discover and MIB-walk Orion will figure out that I have multiple IP addrs -- which would show up in the node details.  Once done I could then select the ip addr that I would like monitored ( like the loopback), and perhaps even designate alternative/secondary IPs for monitoring .

                 

                • Re: If you're curious as to what we're working on...

                  Thank you for giving us some insight on what we can expect in the future. 

                  What about allowing Orion NPM to differentiate between systems based on the SNMP port?  I know I've seen a few posts where people are trying to monitor modules in chassis-based systems with a single management IP or (as in my case) using PAT to traverse a firewall when provisioning VPNs is not a solution.  This functionality seems like it is a high priority for those users attempting to make it happen.

                  Also, is there a place that describes the upgrade cycles for NPM?  If I have to say "the feature might be coming," we'll want to know how long it could take.  For instance, if it doesn't make it in the next upgrade cycle, how long will I have to wait for the next cycle for a possible update? 

                   

                  Thanks for all of the information.

                  • Re: If you're curious as to what we're working on...
                    chrissmail

                    Hi there,

                    Would it be possible to add the ability to assign thresholds specifically to disk volumes and also to memory and CPU? This would be a big win with my company!

                    Thanks,

                    Chris.

                      • Re: If you're curious as to what we're working on...
                        chrissmail

                        Also it would be really cool if you could have the ability to project future trends. The graphs could say (and notify in an alert) that a volume will run out of space in X number of days.

                          • Re: If you're curious as to what we're working on...
                            chrissmail

                            Also smarter email routing will be good. Currently we use email rules to forward email alerts to the appropriate people based on the content of the message. I would like this to be part of the product rather than having to use email rules. Possibly having the ability to have variables in the 'To' field of the email alert would be useful.

                              • Re: If you're curious as to what we're working on...
                                bshopp

                                Thanks for the responses.  I wanted to address each comment

                                Spiky - this is on our list of items to add to the product, but based on current priority did not make the cut for the next release

                                Jason - this item is on our list as well, but as indicated in the post, did not make the cut for the next release.  Regarding your second question it is hard to answer that.  Mainly because our priorities today can and does change over time based on feedback from ya'll in the community, internal and external factors.  So what is high on my list today may move down the list based on these events.  Hopefully that makes sense.  The last thing I want to do is commit something until I know it is a sure thing which means we are working on it.

                                Chris - we do have the first item in the system for a future release.  Regarding trending you can do some of that today with 9.5, if you edit a chart and set the period for sometime in the future for example, you will see a trend line out into that future time based on the historical data we have gathered.  What is not there is alerting on that, which is something we have heard from other users as well.  And finally regarding the email item, it is something I will log into the system, but have not heard a ton about from ya'll.  So for those reading this, if this or any of these things are important to you, please chime in and tell us so we can shift priorities based on the feedback.

                                kupjones - are you seeing this with 9.5?  I beleive in that case we'd pick one of the IP's and only use that.  If this is not the case, please let us know.

                                  • Re: If you're curious as to what we're working on...

                                    Brandon, sorry I was not clear -- so I cobbled together a screen shot of NPM with what I *think* the Presentation should look like.  Operationally, this is how I think it should work:

                                    - SNMP will give you both the interface table (L2) and the IP table (L3). 

                                    - We are interested in both. 

                                    -- The L2 table is good as it allows me to alert on events that occur at Layer 2.

                                    -- What is missing is a really good representation of L3.  If I want to monitor a Layer 3 I must create a separate device for each IP addr -- and thus I end up with multiple devices, all with the exact same L2 interfaces.  I should be able to discover a device via a seed IP, then "tell" Orion which L2/L3 combinations I am interested in.  I attached the cobbled together "screenshot" of NPM where I show not only the monitored physical interfaces but also the IP addresses and the L2 interface that they are attached to.  The L3 addresses should be selectable as to whether or not they are being monitored.

                                    One big reason for having multiple L3's is if reachability is achieved through multiple L3 routes.  Thus, I want to be notified when an L2 goes down but it is equally important for me to know that the *device* is still reachable via the alternative L3 interface.

                                    I believe NPM 9.5 still has the default behavior of "picking" an IP address, unless I specifically discover an address.

                                    Does this help?

                                    • Re: If you're curious as to what we're working on...
                                      Seashore

                                      Hi

                                      Yes, putting variables in the "To" and "server" feilds of alerts would be realy nice to have.

                                      Changed email server a few weeks ago and there where a lot of rules to change in...

                                       

                                      Thanks for a good work!

                                        • Re: If you're curious as to what we're working on...
                                          bshopp

                                          If you are a current customer under active maintenance and are interested in helping us test out a beta of Connect Now and VMWare API support, please send me a PM via thwack.  This is on a first come first serve basis, so we may not be able to have everyone participate.  In your note to me, please include your SWID and which of the two features you are interested in helping us with. 

                                      • Re: If you're curious as to what we're working on...

                                        allowing these e-mail rules to be built into Orion would be great.  This would be a benefit for our environment as well.

                                    • Re: If you're curious as to what we're working on...
                                      KenKasmar

                                      Chis, the way that I have worked around the issue of limited threshold controls is to add a number of custom properties that I use as 'Alert Package Designators'.  These are organized in a tiered structure with system wide defauts at the bottom of the list and the most granular at the top (volume & interface specific).  Then I build alert packages for the different types of alerting that I want to do and populate the appropriate custom property with the package codes (example: DISKMSSQL001 for one of my disk volume alerting packages used for MS SQL servers).

                                      This seems to allow for a fair amount of flexibility to those managing servers while keeping things grouped up enough (and query-able) to be able to maintain things.

                                    • Re: If you're curious as to what we're working on...
                                      kbaumann


                                       

                                      PLEASE NOTE:  We are working on these items based on this priority order, but this is NOT a commitment that all of these enhancements will make the next release.  We are working on a number of other smaller features in parallel.   If you have comments or questions on any of these items (e.g. how would it work?) or would like to be included in a preview demo, please let us know!

                                       





                                      How about native support for Juniper/Netscreen devices? This is a huge issue for me and for others based on this forum post:

                                      Re: Juniper Netscreen sub-interfaces? 

                                      • Re: If you're curious as to what we're working on...
                                        khelgoth

                                        I'd be happy if I could filter "last 25 events" and "alerts" on the "Network Summary" homepage to show useful things that I want to know. i.e. filter out things that I don't want...to wit:  I don't need to see every interface in my 2600+ node environment changing speed, going up, going down, etc. - it hides the important events in an avalanche of general info.

                                         I've searched the forums, but there doesn't seem to be much support in NPM 9.5 to do this.

                                        Cheers,
                                        Kurt

                                        • Re: If you're curious as to what we're working on...
                                          chuco

                                          Can you guys please add a recurring maintenance schedule to NPM for monitoring. This will be something that all of us would like to see I am sure. We have maintenance night every Thursday and we just hate receiving 183 emails of spam because of reboots. I've asked about this many of times and I get the the typical we are working on it. We have owned Solarwinds for a while now and we haven't seen it yet as a feature in the NPM 9 versions.

                                          • Re: If you're curious as to what we're working on...
                                            extrands

                                            How about more granularity on interfaces added during discovery, perhaps the ability to limit interfaces to only those that have a CDP neighbor?  On a large network, you can end up with thousands of interfaces you don't necessarily want to monitor, and it can be a time consuming process to remove them.  Adding CDP neighbor to the import settings, in in addtion to the current operationally up, operationally down and shutdown would be great. 

                                            Some logic that knows a device has been previously discovered, and asks, not assumes, whether a newly discovered operationally up interface should be added would be nice too.  If you go to scheduled discovery, a great addition, we need some additional ways to filter what ends up being added.

                                              • Re: If you're curious as to what we're working on...
                                                Miron

                                                Hi,

                                                I would have liked to see some development into the following few items among others.

                                                Network Atlas:

                                                • Being able to resize labels/text boxes
                                                • Position Securing of Labels/text boxes which dont move when other objects are added
                                                • Ability to insert a table
                                                • Automatic addition/population of table by possible object attibutes.
                                                • Ability to drag and drop objects into a table cell
                                                • Increased icon support
                                                • Generic objects such as racks of different sizes, equipment of different sizes for quick creation
                                                • automatic/dynamic map creation based on object properties. Ie this map has all objects which have custom property: Priority - High.
                                                • Option to automatically delete all presence of nodes on map if you delete a node rather than going through maps individually.

                                                Some focus on Services in addition to individual nodes/applications.

                                                • The ability to create a service object
                                                • ability to assign nodes/interfaces/applications/custom to the service/s
                                                • Define attributes of the service such as sla,kpi,owner,1st line support,base service,
                                                • be able to create a resource page per Service which can serve as a service catalogue description including service dependencies(like a service catalogue entry)
                                                • ability to add service objects as dependant on other service objects
                                                • ability to alert on service object status

                                                 

                                                Web Portal

                                                • Inbuilt ability for map rotation
                                                • drag/resize/reposition resources with web portal for end users and saved per user.

                                                User Management

                                                • Role based access
                                                • creation of groups
                                                • mapping into AD/radius/tacacs  authentication
                                                • change management options
                                                • options to have approval process
                                                • logging user changes

                                                 

                                                Reporting

                                                • The option to keep past node statistics for accurate reporting.Currently If you remove a node your reporting metrics will change and no longer be accurate
                                                • Would be good to report on the trends on number of nodes/applications/interface managed so that you can show how the server performs compared to number of monitored objects
                                                • all of the other request that fellow users have had on here like addition of graphs

                                                Health

                                                • better support for inbuilt ability to automatically recover from service failure (like the current npm process having to be manually restarted after polling engine stops polling
                                                • or better alert mechanism to inform users of server issues.

                                                 

                                                Regards

                                                 

                                                Miron

                                                  • Re: If you're curious as to what we're working on...
                                                    bshopp

                                                    Good list, so I am going to respond to each inline here.  Please feel free to PM me offline to discuss these more in depth further.

                                                    Network Atlas:

                                                    • Being able to resize labels/text boxes
                                                      So just an FYI, if you want to do multi-line you can do a shift-enter to get a new line. 
                                                    • Position Securing of Labels/text boxes which dont move when other objects are added
                                                      So other customers who have hit this issue have used the linked backgrounds feature and places the image in the Orion web server directory and pointed at it to get around this. 
                                                    • Ability to insert a table
                                                      Can you define this a little further.  What are you looking to do?  What type of data? 
                                                    • Automatic addition/population of table by possible object attibutes.
                                                      Again, please explain a little further. 
                                                    • Ability to drag and drop objects into a table cell
                                                      I think this will make more sense to me when I get the other data from above 
                                                    • Increased icon support
                                                      Fair enough, there is another thread of thwack where our UI designer is asking for feedback 
                                                    • Generic objects such as racks of different sizes, equipment of different sizes for quick creation
                                                    • automatic/dynamic map creation based on object properties. Ie this map has all objects which have custom property: Priority - High.
                                                      Understood.  Connect Now which I have outlined above is the first step, but I completely agree with you 
                                                    • Option to automatically delete all presence of nodes on map if you delete a node rather than going through maps individually.
                                                      Understand the use case 

                                                    Some focus on Services in addition to individual nodes/applications.

                                                    • The ability to create a service object
                                                    • ability to assign nodes/interfaces/applications/custom to the service/s
                                                    • Define attributes of the service such as sla,kpi,owner,1st line support,base service,
                                                    • be able to create a resource page per Service which can serve as a service catalogue description including service dependencies(like a service catalogue entry)
                                                    • ability to add service objects as dependant on other service objects
                                                    • ability to alert on service object status
                                                      So here you require the ability to group together one or more services or apps together into a Service for example.  So Service A contains app A, app B, service C, nodes A,B,C 

                                                     

                                                    Web Portal

                                                    • Inbuilt ability for map rotation
                                                      What are you using today?  Alot of users use the Firefox plugin, does this not work? 
                                                    • drag/resize/reposition resources with web portal for end users and saved per user.
                                                       

                                                    User Management

                                                    • Role based access
                                                    • creation of groups
                                                    • mapping into AD/radius/tacacs  authentication
                                                    • change management options
                                                      What type of change management options? 
                                                    • options to have approval process
                                                      What types of actions do you want approval around? 
                                                    • logging user changes
                                                      We have this logged as audit trail 

                                                     

                                                    Reporting

                                                    • The option to keep past node statistics for accurate reporting.Currently If you remove a node your reporting metrics will change and no longer be accurate
                                                    • Would be good to report on the trends on number of nodes/applications/interface managed so that you can show how the server performs compared to number of monitored objects
                                                    • all of the other request that fellow users have had on here like addition of graphs

                                                    Health

                                                    • better support for inbuilt ability to automatically recover from service failure (like the current npm process having to be manually restarted after polling engine stops polling 
                                                    • or better alert mechanism to inform users of server issues.
                                                      Fair enough 

                                                      • Re: If you're curious as to what we're working on...
                                                        Miron

                                                        Hi,

                                                        I will expand and add further detail to the items listed and commented on:

                                                        Network Atlas:

                                                        • Being able to resize labels/text boxes
                                                          At the moment there is not much different between a text box and label. You used to be to drag the handles of the label to create a label of the desired size regardless of quanity of text. So if you want to have a series of labels back to back to each other and want them to share size characteristics you are no longer able to.
                                                        • Position Securing of Labels/text boxes which dont move when other objects are added
                                                          So other customers who have hit this issue have used the linked backgrounds feature and places the image in the Orion web server directory and pointed at it to get around this. Not sure you should require a work around, also if you decide to change a background depending on the objects in the map you have to reposition them as they move around instead of staying put
                                                        • Ability to insert a table
                                                          Can you define this a little further.  What are you looking to do?  What type of data?  Well imagine you want to create a table which lists all the urls that you monitor in the 1st column and the status in the second. Now to do this currently you would have to drag the application monitor onto the map for each url checked, assign a graphic and possible alter the label. You would then have to create a table in whatever program you choose (for instance powerpoint) then save this as an image and import it as a background into network atlas. At this point you would have to put all the URL Application monitors into the right hand column and adjust the size to fit the background , and position all the labels in column 1.   Now this might work and look okay in network atlas but when you display it through the web portal you can get inconsistences in the rendering of the table and what looks like crisp edges in network atlas dont quite look right for user. You would also have the list at the bottom.                                                                     1) Instead imagine if you had the ability to create a table or draw a table. You could right click column one and say that its properties is label, right click the second column and select its properties as status. Then you drag an object into the row and the label in column 1 is filled in automaticaly and column 2 becomes a square with the status colour of the object.                                                                                                     2) Same as number 1 however instead of dragging the objects and adding more rows if you need to you are able to add rules to the table. So you are able to say that column 1 will contain lables of all nodes with specific criteria - basic filter (so any application that has url monitor). Then if you create or remove url application monitors they automatically get added/removed from the table and the table resizes acccordingly                                                                                                                                                      3) Now if this was possible great but lets make it cooler. Lets say we want to add columns for SLA and availabality. So the SLA column takes its content from a custom propert like 95%, in the available column you are able to show the current/weekly availablity. To make it tricker is could also change colour to indicate if it is below the SLA based on alert thresholds for a device or group of devices.
                                                        • Automatic addition/population of table by possible object attibutes.
                                                          Again, please explain a little further.  See Above
                                                        • Ability to drag and drop objects into a table cell
                                                          I think this will make more sense to me when I get the other data from above  . See Above
                                                        • Increased icon support
                                                          Fair enough, there is another thread of thwack where our UI designer is asking for feedback 
                                                        • Generic objects such as racks of different sizes, equipment of different sizes for quick creation.  To expand on this a snap in feature would be good. So you drag a 1U server into U42 and it snaps in.
                                                        • automatic/dynamic map creation based on object properties. Ie this map has all objects which have custom property: Priority - High.
                                                          Understood.  Connect Now which I have outlined above is the first step, but I completely agree with you 
                                                        • Option to automatically delete all presence of nodes on map if you delete a node rather than going through maps individually.
                                                          Understand the use case 
                                                        • NEW: ability create shapes and objects with network atlas, lets say i want to draw a circle or create a box to go around some nodes, these shapes can have the possiblity of having nodes assigned to them like normal objects.  

                                                        Some focus on Services in addition to individual nodes/applications.

                                                        • The ability to create a service object
                                                        • ability to assign nodes/interfaces/applications/custom to the service/s
                                                        • Define attributes of the service such as sla,kpi,owner,1st line support,base service,
                                                        • be able to create a resource page per Service which can serve as a service catalogue description including service dependencies(like a service catalogue entry)
                                                        • ability to add service objects as dependant on other service objects
                                                        • ability to alert on service object status
                                                          So here you require the ability to group together one or more services or apps together into a Service for example.  So Service A contains app A, app B, service C, nodes A,B,C  - Correct Lets take for example one of my wireless services. Wireless SSID X - It uses Thin wireless devices, WLC, radius server, active directory and a ASA. Now if one of the access points goes down its not to much of a big deal, but if the radius server or asa is unavailable the service can no longer work.

                                                         

                                                        Web Portal

                                                        • Inbuilt ability for map rotation
                                                          What are you using today?  Alot of users use the Firefox plugin, does this not work?  The firefox plugin works fine, however users have the option to use browsers of their choice, so if we want them to have rotating summary screens it would be good if it was browser independent
                                                        • drag/resize/reposition resources with web portal for end users and saved per user. Kind of like he google portal where the user can reposition resources, image also if you wanted to just detach a resource to view in its own window.
                                                           

                                                        Maintenance Windows (Newly Added)

                                                        • I see this has been talked about already but again lets take it a step further, lets say you have a particular time of the year when a certain app is critical, so lets say for the month of july every year we want to increase the polling times on this from once in 5 min to one in 2 min, and increase its custom attribute Priority to High. We want the Priority to High because we have a general alert that will send out text messages only for applications/nodes with a high priority.(examples of applications that are time dependent payroll systems are mostly important the week before payday, other times of the month if they go down it is not worth someone getting out of bed for.
                                                        • How would this look, well lets imagine a microsoft calendar type scheduling assistant /rules wizard. We  select the period or recurrence, we create a new rule for this within this rule we define a filter/s to address the target objects, we define a set of actions that can change the contents of a custom property or tag, we can define polling time, we can define managed status. We may need multiple rules if we have overlapping change events. So one central place to do of it rather than having to do it by object individually. It would also give you one place to view all the changes away from the norm in one place. Obviously the changes defined would be reversed one that occurence was over.

                                                        User Management

                                                        • Role based access
                                                        • creation of groups
                                                        • mapping into AD/radius/tacacs  authentication
                                                        • change management options
                                                          What type of change management options? Lets say someone is going to edit something on a node. Administrator requires that an audit trail (not sure this is what you mean below)of changes to devices also needs to be maintained. So that if you look at change history for device you can see that its name was changed, had an interface unmanaged, and snmp changed. When someone does the change either they are automatically assigned a job number for the change which can be tied back to formal change management process or they can input the corporate change id number.
                                                        • options to have approval process
                                                          What types of actions do you want approval around?  If a user is going to change a node of a certain priority they do they change but it doesnt take effect until someone with approval access to that level authorise the change, this would extend to network atlas where you have edit permissions much like windows shares.
                                                        • logging user changes
                                                          We have this logged as audit trail 

                                                         

                                                        Reporting

                                                        • The option to keep past node statistics for accurate reporting.Currently If you remove a node your reporting metrics will change and no longer be accurate
                                                        • Would be good to report on the trends on number of nodes/applications/interface managed so that you can show how the server performs compared to number of monitored objects
                                                        • all of the other request that fellow users have had on here like addition of graphs

                                                        Health

                                                        • better support for inbuilt ability to automatically recover from service failure (like the current npm process having to be manually restarted after polling engine stops polling 
                                                        • or better alert mechanism to inform users of server issues.
                                                          Fair enough 

                                                         

                                                          • Re: If you're curious as to what we're working on...
                                                            SLXer

                                                            Miron, what a thoughtful list.. very nice

                                                            • Re: If you're curious as to what we're working on...


                                                               

                                                               

                                                              Network Atlas:

                                                              • Being able to resize labels/text boxes
                                                                At the moment there is not much different between a text box and label. You used to be to drag the handles of the label to create a label of the desired size regardless of quanity of text. So if you want to have a series of labels back to back to each other and want them to share size characteristics you are no longer able to.

                                                               



                                                              Is there a change coming to make a distinction between a text box and a label? If they are identical, is one being phased out in favor of the other?

                                                              I'm in the process of updating all our network maps at the moment...

                                                               

                                                              Thanks,

                                                              -Joe

                                                      • Re: If you're curious as to what we're working on...

                                                        If you're looking for suggestions, how about this:

                                                         

                                                        Why not have predefined "Pollers" for platforms, such as Cisco, VMware, etc... Much akin to Microsoft Operations Manager's "Management Packs"? I like the quick turn up of an NPM install, but I wish there were more data available in a quick and easy fashion. It seems that just a bit more info that could be available, like CDP neighbors, etc. would be a quick win and great value proposition to your product.

                                                        • Re: If you're curious as to what we're working on...
                                                          whitejcdc

                                                          Couple of small things I would like to see done...

                                                          1) In the Admin section of the website,  I continually have to go back in to adjust the Limitation for a acct.  Like say for a network eng acct, they need all Ciscos, but everytime we add a new device type the limitation has to be changed.  Would it be possible for keyword matching on type like "Cisco" or something like that?

                                                          2) For unmanaged time set on nodes, it would be nice to see exactly the time frame its been set for on the node details page somewhere.

                                                          • Re: If you're curious as to what we're working on...

                                                            One other suggestion that is becoming a serious problem in my environment is the ability to monitor multiple contexts, and multiple devices via the same IP.

                                                            • Re: If you're curious as to what we're working on...

                                                              I don't know if this has been mentioned, but we would like to have more information in the graphs.  Some of the items we have troubles with for graphs that provide a point of view over a longer time period are:

                                                              • It's hard to identify the specific day or time of day that an event in the graph occured.
                                                              • Information such as the Maximum, Minimum, and average values are not included in the graphs, specifically the APM Statistic graphs.
                                                              • Sometimes the text on the graph overlaps, which ends up looking messy.
                                                              • The 95th percentile lines for interfaces (one for transmit and the other for recieved bandwidth usage) are styled exactly the same.  This means that interfaces with similar transmit and receive values have 95th percentile lines that cannot be visually related to one piece of data over the other.
                                                              • While not nearly as significant as the points above, it would be great to have a mouseover option on the graph, as is done in other systems such as ESX.  When you mouse over a line, it shows you the specific data for that point in the graph including the time and the value.

                                                              I know some of the information, such as min. vs. max values can be retreived by choosing "raw data" or through the use of other web parts, but it would be best to have that information on the graph, where it remains relevant to our users.

                                                              • Re: If you're curious as to what we're working on...
                                                                gbrance

                                                                So what are you guys doing for SQL Cluster support?  We continually are having issues during failover with Orion stops writing to the database and then it doesn't start writing again after the failover is complete.  Wind up rebooting both poller engines and then it will start writing again.  Maybe build some logic into Orion where it will try to write again every 15mins instead of giving up after 5min and then never trying again. 

                                                                • Re: If you're curious as to what we're working on...
                                                                  smartd

                                                                  While i understand that Cisco has a large market share, Juniper has made significant inroads to enterprise networks.  With OEM agreements with Dell and IBM, you can predict this increasing.  While we have almost 1000 cisco routers, we chose to bypass Cisco Nexus and rebuild our data centers with Juniper gear.  I'd like to see a little more focus on Juniper equipment compatibility.

                                                                  -=Dan=-

                                                                    • Re: If you're curious as to what we're working on...

                                                                      I hope that decision works out - sadly, I'm not sure Juniper gets it.  I just got through reading an interview (http://www.networkworld.com/news/2010/020810-juniper-stratus.html?ts0hb&story=stratus) on the two Juniper visionaries and frankly their explanation of what they believe drives a datacenter is odd and the timeframe for new Stratus product is really odd -- especially given that they just got through bashing Cisco for the "gradual" release of Nexus.

                                                                      Ultimately, customers drive the direction of any product and SW is no exception.  Witness this thread!

                                                                        • Re: If you're curious as to what we're working on...
                                                                          smartd

                                                                          Cisco's decision to push FCOE so hard with us and goal to crush Brocade, the fact that Nexus was not developed under the Cisco wing so little non-Nexus integration to the rest of the Cisco products, and their decision shoot at blade center manufacturers and systems support folk with their new blade server and virtual switch (NX-1000) components that poise Systems against Networking groups drove us away. 

                                                                          Stratus is the goal of clustering all datacenter core switches into a single virtual cluster with ultra highspeed cluster cabling between EX8200 cores.  Sound more interesting than running a "Virtual System" partitioned linux OS based Nexus switch.

                                                                          -=Dan=-

                                                                      • Re: If you're curious as to what we're working on...

                                                                        I know it has been mentioned in the past, but we'd like to graph the polling completion percentage.  Seeing other details on the Orion web page regarding the polling engine would be an added bonus.  We've made some recent changes to our system that I'm hoping will help us maintain a stable completion rate above 90%.  We're relying on some quick glances and performance metrics from the database server to suggest the results.  Graphing this would help us know for sure.

                                                                        • Re: If you're curious as to what we're working on...
                                                                          SLXer

                                                                          I am sure it has been said before.. but ill mention it again.. being able to graph multiple sources on a single graph is important. Has there been any discussion on this? Im thankful that solarwinds allows me access to the raw data so that i can generate these graphs with ms sql but frankly the less i need to work with a dba the happier i am.

                                                                           

                                                                          simple case in point... lets say your monitoring 3 or more seperate internet connections with both npm and apm to validate resource performance. Each gateway has 6 http 6 https 3 dns prime 3 dns secondary monitored.. Individual graphs are difficult to use in determining if a performance variation is do to bandwidth constriction on the pipe or on the internet. It is necessary to view each target group on one graph that illustrates general target latency and application latency together.

                                                                          On a separate note. I might suggest (though i doubt it will ever be considered) a mechanism for managing routing on NPM. In order to do the tricky monitoring im doing it unfortunately requires a whole lot of dancing with routes on the local server in order to send the right traffic out the right interface.

                                                                          How is everyone loving 9.5.1 =)

                                                                          • Re: If you're curious as to what we're working on...

                                                                            Hi Bshopp thank you for your information. This is a quite nice posting about what you are doing recently. From this post people is informed about your activity. So from my point of view this is an useful post. That is it. Thank you.

                                                                            • Re: If you're curious as to what we're working on...

                                                                              Solarwinds still can't monitor Cisco SPAN ports. It would be nice to be able to force NPM to record these port's statistics. 

                                                                              • Re: If you're curious as to what we're working on...
                                                                                ecklerwr1

                                                                                Quick Q&A clarification on enhancements:

                                                                                Syslog and Traps –

                                                                                Ability to change the status of an interface in Orion NPM based on Trap message

                                                                                Does this mean changing the status of interface to down (red) when a trap down message is received (before polling figures it out) or is it the ability to do something like... if I received a trap from firewall on for example a DOS attack on a network segment... then have the ability to change an interface into Admin Down status for example to do something like automate the cuting off of the DOS attack at the source?

                                                                              • Connect Now - the ability to drag one or more nodes on to a map and based on the last discovery topological data, automatically connect those device together on the map
                                                                              • Does this mean network atlas could automatically create a line between nodes because it knows the interfaces are connected for example on a WAN circuit between two nodes?  And if this is this case would the status of the line then be dependent on the interfaces between nodes for status without you having to drag the interfaces into the line properties window?  Maybe I'm missunderstanding what this enhancement is???

                                                                                I don't mean to press I just don't quite understand the description on these two... I'm sure a lot of work has gone into trying to get the Vsphere 4 support working... I hope VMWARE documented their new API well :}

                                                                                  • Re: If you're curious as to what we're working on...
                                                                                    bshopp

                                                                                    #1 - mean changing the status of interface to down (red) when a trap down message is received (before polling figures it out)

                                                                                    #2 - This is the ability to drag one or more nodes onto a map and we automatically connect them based on discovery data.  And yes to your second part on status

                                                                                    Feel free to PM me via thwack and we can do a GoTo

                                                                                  • Re: If you're curious as to what we're working on...
                                                                                    twitten

                                                                                    What about being able to create a list of "interesting" interfaces based on device type and automatically add only those interfaces for monitoring? Also, what about being able to periodically re-evaluate this list and add any new interfaces that match the "interested" list?

                                                                                    • Re: If you're curious as to what we're working on...
                                                                                      SLXer

                                                                                      Will complex scenarios be supported or is this a one for one relationship between a single syslog message received and an identified interface? ie. syslog message containing X from devices A & B and syslog message containing Y on device C received within a specified time frame will result in a positive trigger on specified interface for a defined period of time..

                                                                                      Is this limited to effecting the status of interfaces? What about nodes or APM components?

                                                                                      • Re: If you're curious as to what we're working on...
                                                                                        SLXer

                                                                                        Any plans to support more than one item on a graph?

                                                                                          • Re: If you're curious as to what we're working on...
                                                                                            BakerD

                                                                                            Sorry if this has been mentioned already as I just skimmed the thread.

                                                                                             

                                                                                            Anyway to add the option to email a page as a pdf.  Say you drill down to a certain point, or get a graph all setup the way you want, then be able to save that view as a pdf and/or email it to someone.

                                                                                            I have Acrobat installed and I can print my browser page to a pdf file, but it looks like ****.  I have to blow it up to 125% to make anything out and then pictures/text etc are still blurry.

                                                                                             

                                                                                            Thanks

                                                                                          • Re: If you're curious as to what we're working on...

                                                                                            Do you know if this will be in the next release? Orion MaxBytes like MRTG This is a must have for me requested awhile ago and many others have this issue as well.

                                                                                            • Re: If you're curious as to what we're working on...
                                                                                              mhh351

                                                                                              Yes, please include me in the preview demo.

                                                                                              • Re: If you're curious as to what we're working on...
                                                                                                SLXer

                                                                                                Here is a quick suggestion. It would be nice if the down node web component had the same tree functionality as the node tree web object. 

                                                                                                • Re: If you're curious as to what we're working on...
                                                                                                  brian@rdu

                                                                                                  Brandon,

                                                                                                  Cisco WCS just upgraded to 6.0.181.0 and appears to be getting more data from the controllers (version 6.0.196.0).  Is there anything coming out soon that might enhance the v10 Wireless side of things?

                                                                                                  • Re: If you're curious as to what we're working on...

                                                                                                    Hello,

                                                                                                    I don't know if this is the good post for that but I would have suggestion about enhancements. It is about reports. The content is OK, but I think presentation should be improved. Actually we just can get a spreadsheet with a lot of numbers. It would be nice to be able to produce automatically graph with these datas, or to put some colors in the table in order to make it easier to read (for example, at least one line (or column) in white, the next in grey, the next in white and so on) in order to make it more readable... I think there is a work to do to make reports more usable, because reports can be very usefull, report writer is a very good idea, but the output, the reports are not "ergonomic" enough to be really usable.

                                                                                                    Thanks,

                                                                                                    Nicolas

                                                                                                    • Re: If you're curious as to what we're working on...

                                                                                                      Other suggestions, about the node details...

                                                                                                      At the beginning I was surprise to see that I have CPU and memory gauges only for Cisco equipments and not for other equipments (BATM, Redback, OneAccess, Ruby, TPLink...). I have found how to create it for each equipment vendor with universal poller (great tool). So now, I have the info for all the equipments, but the issue is now in the top ten. The top ten include only Cisco equipments and we have to create a different top ten for each kind of equipments. This means 6 Top ten for CPU, 6 Top ten for memory etc... Like this, Top ten is no more really interesting. It would be really interesting to be able to create a top ten (or modify an existing) with datas taken from multiple pollers. I think this development is necessary to make NPM multi-vendor capable.

                                                                                                      Thank you,

                                                                                                      Nicolas

                                                                                                      • Re: If you're curious as to what we're working on...
                                                                                                        mhh351

                                                                                                        I would like to reiterate that I would like the function of NO Discovery on the start up of the system. Just start polling. If the database exists, don't do anything more, jus tstart loading and polling. (Of course you could check off to have it do this if you wanted.)

                                                                                                        Also, could you finally give the poller the ability to "Rediscover" by day and time?  Right now it is only possible to give the amount of time between intervals. If the poller is restarted, THAT is the time that is referenced for the next rediscovery. That can happen during the busiest time of the business week/day and Orion just slows down for about 2 hours.

                                                                                                         

                                                                                                        Thanks for the ear.

                                                                                                          • Re: If you're curious as to what we're working on...
                                                                                                            bshopp

                                                                                                            So for the last three posts

                                                                                                            Reports - this is for sure on our list to make reporting better going forward

                                                                                                            CPU/Mem - this works for many vendors out of this box besides Cisco, however we find many vendors have their own proprietary MIB's where they store this info so the standard place we look they do not support.  We will add them, but we base this in priority from ya'll

                                                                                                            On this last post, please define rediscovery, do you mean network sonar or the node polling rediscovery?

                                                                                                              • Re: If you're curious as to what we're working on...
                                                                                                                mhh351

                                                                                                                Thank you for the reply.

                                                                                                                This would be for NPM and its polling engine. There are really two issues here.

                                                                                                                1. After the poller finally loads into memory, it proceeds to rediscover the entire known database. That is not needed if the database is intact. This process can take up to one hour on the start up of Netperfmon. This gets in the way of actual polling for status and statistics. If one reads the events on the main default web page while the system is coming back up, one sees very strange messages going  by about nodes down and up while the monitored devices and the process is rediscovering and the database is updated.

                                                                                                                I would prefer to make the Polling engines more efficient by allowing for NO start up re-discovery.

                                                                                                                2. Since this also sets the time of the next rediscovery, we have no way of telling the polling engine to only do rediscovery on a set day and time. (Example:  I want to rediscover all devices only on Saturday at 11:00 p.m. each week    or   every day at 10:30 pm, etc.)

                                                                                                                Today, I have to set the number of minutes for a weekly rediscovery, 10080, but I cannot tell the system when the start time should be.

                                                                                                                I hope that this is of help and that I explained it correctly.

                                                                                                                Thanks for the time,

                                                                                                                Mark

                                                                                                                • Re: If you're curious as to what we're working on...
                                                                                                                  smartd

                                                                                                                  Hey Brandon,

                                                                                                                  You guys should look at the feature voting service at uservoice.com.  I use one with Yammer at http://yammer.uservoice.com/forums/22714-general-feedback

                                                                                                                  Great for adding suggestions, then voting on the suggestions.

                                                                                                              • Re: If you're curious as to what we're working on...
                                                                                                                SLXer

                                                                                                                Searchable events is a long overdue option. The ability to do keyword searches for events within a specific time frame is has real value.

                                                                                                                • Re: If you're curious as to what we're working on...
                                                                                                                  SLXer

                                                                                                                      

                                                                                                                   

                                                                                                                  Like any other large organizations we rely on a number of applications to provide statistical analysis and general performance information. One of these applications has a feature that I think you would be very wise to copy.

                                                                                                                  On every single page there is a link on the footer to request a feature or report a problem. If you employed something like this it would not only allow you to increase the amount of feedback you got but have it be automatically organized based on where it was being generated from..

                                                                                                                  Just seems like a really bright idea that can’t be copyrighted.

                                                                                                                  • Re: If you're curious as to what we're working on...
                                                                                                                    mkomeara

                                                                                                                    We need to create a NOC type of display and when projected or shown on a large display most elements in a view aren''t large enough to be readable from a distance.  Maps can be handled by creating a large map and using larger icons and text, but other resources don't give us that flexibility.  If I set the column width to 900 pixels, the heading, graph and text are the same size as when the column width is set to 400.  When assigning a Resource to a View, it would be great to have one or more of these three choices;

                                                                                                                    1. A checkbox option of scaling everything in a column (like zoom) as the pixel width is increased or decreased (my favorite).

                                                                                                                    2. The percentage over or under 100% to scale everything in a column or in an individual resource.

                                                                                                                    3. The ability to set sizes for individual elements (heading, text, graph, etc.) in a Resource.

                                                                                                                      • Re: If you're curious as to what we're working on...
                                                                                                                        tdanner

                                                                                                                        Why not just use the browser's page zoom feature?

                                                                                                                          • Re: If you're curious as to what we're working on...
                                                                                                                            mkomeara

                                                                                                                            It would solve the problem until Solarwinds adds that functionality to Orion's web portal, but not without challenges;

                                                                                                                            - It enlarges everything on the page, so the heading bar takes up more space on a limited display.

                                                                                                                            - We have four tabs on the IE display that it rotates through and it doesn't seem to be possible to automate the zoom level setting on a per-tab basis.  When IE is closed and opened, everything either changes back to 100% or goes to the new zoom level, depending on the check box in the Advanced tab of Internet Options.  This will be a system that starts the browser automatically and shouldn't require operator intervention, so unless someone knows how to permanently set the zoom level on a per-tab basis, it won't help.

                                                                                                                        • Re: If you're curious as to what we're working on...
                                                                                                                          dclick

                                                                                                                          With all the planned changes to Syslog/Traps, does this mean your moving this into the basic/advanced alerts editor, or we still going to have 2 different alert managers/edtiors ?  I applaud the effort - these are things that are MUCH desired.

                                                                                                                           

                                                                                                                          I am also glad to see work on Network Atlas/ConnectNow! I think these are 2 very under-served parts of the Orion line.  This could be VERY VERY valuable troubleshooting tools, if there was some development time spent on making them more useful.