Issues I've Encountered...
Ok, first of all our company just purchased opmanager to monitor a total of around 250 workstations and 500 servers, plus 50 or so network devices.
While configuring opmanager to monitor a meer 100 servers and 2 network devices, without anything really unusual I've encountered many things that were either far harder than they should have been, hidden and hard to find, or various other things I would just like to list out in the hopes of seeing them improve or be fixed in the next version. I have documented every single "issue" I've encountered from installing it to its current state (maybe 10% done for our company). Think of me as a real life non-biased QA tester ;)
I'm sure at least a few of these things opmanager *can* currently do, but I couldn't find it, so in my eyes it couldn't do it
I'm sure some of these things can be done by tweaking the config files, but I'm a firm believer in keeping software clean, and that means I need to stay out of files like this in a production environment, those changes require documentation, duplication upon upgrades, etc. Some of these are new features, others new ways of doing things, some even bugs.
Here is my current list, but I'm sure I'll add more later ;) Let me know if you have any questions on what I'm talkin about exactly, I'll try to elaborate.
-When adding a device, if its not found change the green INFO bubble to red or yellow, not green
-Have the device added screen return its index to the device name box when it refreshes after adding a device
-When you add a device, it creates a network map with that IP, but the network discovery never shows the difference between that 1 IP thats in the map and a network discovery performed on the entire subnet
-Put servers into groups OTHER than network maps, for business units (production, development, etc)
-Discover a network from an Active Directory OU instead of just an IP Range
-Discover a network from a list of IP/Computernames, instead of just an IP range
-Adding a device needs to accept a list, not just 1 device at a time
-Allow defining of further groups on the home screen (for adding things like file servers, dns server, web servers, etc)
-Support for multiple SNMP strings
-Change the size of the columns in the maps page, some longer entries get truncated
-Customize the map screen to show the services we actually *want* to see, not hardcoded defaults
-When discovering a device, also look at windows services and add any that are setup to be monitored
-When associating services to a device, it doesn't always show all the right devices
-Network map views the subnet needs to allow the details view the main page allows
-Change attributes on multiple machines at once (like changing 100 "unknowns" to "windows 2003" server types)
-Instead of editing all the \conf\ files, hows about a page or 2 that lets you edit them instead of chancing screwing up config files.
-Hows about installing a shortcut to stop and start opmanager, its a pain to go to services every time to do that.
-Option for "rediscover all" instead of having to click on each subnet
-Redscovering a subnet needs an option to NOT add new devices, sometimes I may only have 1-2 devices out of 100 that I want discovered.
-Option when discovering to drop DNS suffixes, show IP only, or pull names from a DNS server to avoid having IPs only.
-Spaces in user names for discovering devices isn't supported
-Multiple IPs for a single device (some windows servers commonly have 5 or so IP, and a different device for each sucks)
-Need multiple dependencies, some computers could be hooked to 2 switches for redundancy as an example
-When doing a discovery it can't tell what microsoft OS is on a machine??? 2k3 is unknown???
-Have multiple drop downs and entries on the device page, so many updates can be done with a single "save updates" instead of having a popup for each change
-Configurable infrastructure Views, we don't need UPS/Wireless/Printers/Firewalls, but we do need VIPs
-Custom WMI monitors (Current connections, Gets/Sec, etc)
-Use a script/application for a monitor where an error level is returned
-Add a comment field to devices, so at a quick glance you can see a comment without going down to additional fields
-An alarm is needed if a poll cannot be conducted (especially from invalid username or SNMP string)
-When adding a service monitor, the checkboxes are way over on the left and its hard to tell which box is associated to which device (move checkboxes to right?)
-Ability to use site from IIS instead of Apache
-Ability to select lots of objects and set thresholds on monitors for them (what if I want to implement a error by CPU being over 95% for 3 polls on 100 devices?!)
-When adding a device, allow the "name" of the device to be added at this point instead of defaulting to whats detected by DNS/SNMP/etc
-When adding a monitor, allow adding multiple monitors at once, its a pain to add 10 WWW monitors to 50 web servers
-Ability to disable just 1 particular monitor on a device (say I wanna stop my web service a few minutes)
-When you have spaces in a service name, it fails the polling and shows as failed (pretty damned critical error IMO!)
-Ability to monitor a scheduled task
-Ability to monitor a folder for total space or # of files
-Configure the icons, be able to drop the device names, change the font color, and shorten the space between icons
-Draw dependencies on the network map page
-No adding of service monitors in quick configuration
-Ability to set an alarm type (especially event logs) for just information on the device screen without creating an alarm on the alarm screen
Of course, if ya really dont care about this let me know so I'll stop documenting it ;)
Eric Smith
match.com
While configuring opmanager to monitor a meer 100 servers and 2 network devices, without anything really unusual I've encountered many things that were either far harder than they should have been, hidden and hard to find, or various other things I would just like to list out in the hopes of seeing them improve or be fixed in the next version. I have documented every single "issue" I've encountered from installing it to its current state (maybe 10% done for our company). Think of me as a real life non-biased QA tester ;)
I'm sure at least a few of these things opmanager *can* currently do, but I couldn't find it, so in my eyes it couldn't do it
I'm sure some of these things can be done by tweaking the config files, but I'm a firm believer in keeping software clean, and that means I need to stay out of files like this in a production environment, those changes require documentation, duplication upon upgrades, etc. Some of these are new features, others new ways of doing things, some even bugs.Here is my current list, but I'm sure I'll add more later ;) Let me know if you have any questions on what I'm talkin about exactly, I'll try to elaborate.
-When adding a device, if its not found change the green INFO bubble to red or yellow, not green
-Have the device added screen return its index to the device name box when it refreshes after adding a device
-When you add a device, it creates a network map with that IP, but the network discovery never shows the difference between that 1 IP thats in the map and a network discovery performed on the entire subnet
-Put servers into groups OTHER than network maps, for business units (production, development, etc)
-Discover a network from an Active Directory OU instead of just an IP Range
-Discover a network from a list of IP/Computernames, instead of just an IP range
-Adding a device needs to accept a list, not just 1 device at a time
-Allow defining of further groups on the home screen (for adding things like file servers, dns server, web servers, etc)
-Support for multiple SNMP strings
-Change the size of the columns in the maps page, some longer entries get truncated
-Customize the map screen to show the services we actually *want* to see, not hardcoded defaults
-When discovering a device, also look at windows services and add any that are setup to be monitored
-When associating services to a device, it doesn't always show all the right devices
-Network map views the subnet needs to allow the details view the main page allows
-Change attributes on multiple machines at once (like changing 100 "unknowns" to "windows 2003" server types)
-Instead of editing all the \conf\ files, hows about a page or 2 that lets you edit them instead of chancing screwing up config files.
-Hows about installing a shortcut to stop and start opmanager, its a pain to go to services every time to do that.
-Option for "rediscover all" instead of having to click on each subnet
-Redscovering a subnet needs an option to NOT add new devices, sometimes I may only have 1-2 devices out of 100 that I want discovered.
-Option when discovering to drop DNS suffixes, show IP only, or pull names from a DNS server to avoid having IPs only.
-Spaces in user names for discovering devices isn't supported
-Multiple IPs for a single device (some windows servers commonly have 5 or so IP, and a different device for each sucks)
-Need multiple dependencies, some computers could be hooked to 2 switches for redundancy as an example
-When doing a discovery it can't tell what microsoft OS is on a machine??? 2k3 is unknown???
-Have multiple drop downs and entries on the device page, so many updates can be done with a single "save updates" instead of having a popup for each change
-Configurable infrastructure Views, we don't need UPS/Wireless/Printers/Firewalls, but we do need VIPs
-Custom WMI monitors (Current connections, Gets/Sec, etc)
-Use a script/application for a monitor where an error level is returned
-Add a comment field to devices, so at a quick glance you can see a comment without going down to additional fields
-An alarm is needed if a poll cannot be conducted (especially from invalid username or SNMP string)
-When adding a service monitor, the checkboxes are way over on the left and its hard to tell which box is associated to which device (move checkboxes to right?)
-Ability to use site from IIS instead of Apache
-Ability to select lots of objects and set thresholds on monitors for them (what if I want to implement a error by CPU being over 95% for 3 polls on 100 devices?!)
-When adding a device, allow the "name" of the device to be added at this point instead of defaulting to whats detected by DNS/SNMP/etc
-When adding a monitor, allow adding multiple monitors at once, its a pain to add 10 WWW monitors to 50 web servers
-Ability to disable just 1 particular monitor on a device (say I wanna stop my web service a few minutes)
-When you have spaces in a service name, it fails the polling and shows as failed (pretty damned critical error IMO!)
-Ability to monitor a scheduled task
-Ability to monitor a folder for total space or # of files
-Configure the icons, be able to drop the device names, change the font color, and shorten the space between icons
-Draw dependencies on the network map page
-No adding of service monitors in quick configuration
-Ability to set an alarm type (especially event logs) for just information on the device screen without creating an alarm on the alarm screen
Of course, if ya really dont care about this let me know so I'll stop documenting it ;)
Eric Smith
match.com