Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

Why do you monitor servers in Orion?

Just to give you some insight into where our heads are at, anytime we evaluate new "application monitoring" capabilities (delivered as a module for Orion), we always examine them through the lens of "application monitoring for the network engineer". For example, this has driven us to focus on adding the following capabilities in the new Orion APM module *currently in beta*:

Network service monitoring (e.g. FTP, HTTP, DNS, etc.)
Generic TCP port monitoring
Process monitoring (same as existing AM module)

All the customers we've spoken to are really happy about where we’re headed. However, we also continue to get requests for deeper server monitoring (e.g. disk I/O, perfmon counters through WMI) and application performance monitoring (e.g. Exchange queues, SQL queries, etc.) in Orion. We have our own beliefs based on our research and past surveys, but I'd like to play dumb for a bit and ask the following questions:

If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?
What other capabilities would you like to see us add around server monitoring and why?
Why would you want to monitor the performance of applications like Exchange, SQL, etc in Orion?
When and how do applications fall into your realm of responsibility as a network engineer?

Find more posts tagged with

Accepted answers

All comments

Network_Guru

If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?
We do not monitor servers with Orion. other than our own NPM servers.
What other capabilities would you like to see us add around server monitoring and why?
Port monitoring, but for routers & switches. This would give us a second method of determining status of the device, besides just ICMP (Is port 22 or 23 available, if so, mark the device as up). TCP is also a more reliable indicator of Network Performance.
Why would you want to monitor the performance of applications like Exchange, SQL, etc in Orion?
Our company has 24,000+ Employees, so IT infrastructure is far more granular than a smaller company. There are other groups using different tools to monitor servers, one for Windows & another for Unix.
When and how do applications fall into your realm of responsibility as a network engineer?
When applications fail, the first thing blamed is always the network. Application monitoring might be good for the network people as another tool to use to absolve the network of any "wrong doing" & help prove it was the application's fault.

hez

1. If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?

We monitor inhouse servers only currently (approx 10 vs 250 network devices), but as we primarily monitor external companies equipment, it would be useful to be able to provide application monitoring as a service also.

2. What other capabilities would you like to see us add around server monitoring and why?

I would say that real application monitoring begins when you do stuff like mom's agents, but thats a bit of Orion's scope, at least for the moment, but the possibility to do custom WMI, and some predefined stuff would be a great start.

3. Why would you want to monitor the performance of applications like Exchange, SQL, etc in Orion?

Knowledge is always good. (maybe not for a good nights sleep, but in general), and to be able to bill more for our services, and as Guru said, being able to shift focus from the network, which is always to blame..

4. When and how do applications fall into your realm of responsibility as a network engineer?

Goes back to the fact that we (as network engineers) very often get the blame, and any tools we then can use to pinpoint problems is a good thing to have.

Another point is that In very many companies around the globe, there isn't a distinction between network engineer and sysadmin, the companies are too small, and with the fast learning curve and implementation of Orion, any added value for the Admin will make em happy.

I Just helped a friend getting his Orion eval up, only took one evening, and he has been self sufficient afterward, and very enthusiastic.

/heZ

dbatten

Network service monitoring (e.g. FTP, HTTP, DNS, etc.)

I hope this includes some robust NTP monitoring. Previously we were using Whatsup for NTP monitoring in our environment. Now that we are using Orion we have augmented NTP using cutstom scripts to create "events" and "alerts" to let us know if there is a problem with NTP service in our environment.

If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?

Primary purpose for our department is to monitor our department (Network Services) servers, and others that create a dependency for our support area. Such as DNS servers--we maintain the DNS service, but the Server team maintains the server.

What other capabilities would you like to see us add around server monitoring and why?

Default and option to create "process sets" with independent monitoring parameters. Instead of selecting each individual process to monitor, I would like to group them together.

e.g. There are several independent Orion processes that run; I would like to group this together to say this is Orion application -- IIS server, NPM, SQL database, etc.

The grouping would allow: grouping of process on same server, and processes across multiple servers.

Why would you want to monitor the performance of applications like Exchange, SQL, etc in Orion?

Smaller organizations may want more robust server & application monitoring, but larger organizations have other support tools and teams for this function.

However, some of the other team members want our department to do performance monitoring for them just for the ease of use of the Orion web interface and graphing capabilities. They do not want us to provide any event or alert capabilities because they have other tools for that purpose. Such as, MOM and Tivoli.

When and how do applications fall into your realm of responsibility as a network engineer?

When the applications we support directly are those that provide the service we manage. Such as, Orion, Cirrus, other network tools, DNS, DHCP, NTP.

In my case, my expertise is actually defining end-to-end (E2E) monitoring recommendations--and specifically providing "application characterization", "network capacity planning" and "network impact assessments. That is, we take packet captures for specific business transactions (multi-tiered environments) and use OPNET's IT Guru products to definitively define the performance E2E; the total time to complete the business transactions. This provides time breakdown for: Server processing, work station processing, network delays, and protocol/congestion, etc.

atsaloli

Hi. I am really glad you are looking for feedback from your clients.

If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?

a. to know when there is an outage so we can fix it (availability monitoring)
b. to keep track of utilization (resource/performance monitoring and capacity planning)

What other capabilities would you like to see us add around server monitoring and why?

I need to be able to monitor the following on my Unix/Linux servers to provide quality support for my environment:

a. CPU utilization

b. swap activity (swapping to and from disk)

c. network interface utilization (% utilization, plus watching for spikes in errors)

d. filesystem utilization(% used, and % i-nodes used)

e. disk device response speed (service times)

f. process monitoring (make sure an application is running)

g. it would also be nice to be able to monitor TCP/IP ports, to make sure that ports that are supposed to be accepting connections, are; and to alert me if any new ports appear that are not specified. (to catch rogue applications/servers.)

Why would you want to monitor the performance of applications like Exchange, SQL, etc in Orion?

I wouldn't.

When and how do applications fall into your realm of responsibility as a network engineer?

They don't. I am responsible for supporting the operating system. Somebody else handles the applications.

Thanks again for asking these questions.

gbrance

At my site we have about 900 users with around 45 physical servers and 30+ virtual servers. We cannot have server administrators and Network administrators, we just don’t have the manpower for this specialization in IT. Personally I am just a LAN administrator and have to be prepared to handle anything from our Dell SAN to a 4507 Catalyst switch. When we first implemented Orion it was primarily to monitor our Network gear but since we didn’t have any monitoring on our Server I decided when I implemented Orion to also monitor our servers because some information was better than no information. Also we did purchase the Application Monitor Module for Orion which allowed us to pull some more information out of Orion. The basic thing about Orion is that we wanted a 1-stop shop for all of our monitoring. Our corporate IT team uses MOM but we found this was cumbersome to use and we just didn’t have the time to work with it, our corporate IT is specialized enough that they can dedicate head count to monitor solutions like Orion. I also found Orion useful to inventory our servers. Using custom properties field you can quickly see how many servers of each model we have and there service tags. Also there is a custom report displayed on the Orion site that I built that shows which is the current active node of our clusters. I even monitor volume size and we get alerts when a volume crosses the 95% threshold. Now that we even have several years of volume data we can even plan volume sizes that will be needed in the future.

Going forward now that Orion 8.x was released and we could grab custom SNMP MIB a lot more in regards to performance data has been opened up in regards to what we monitor. I use this 3^rd party agent from http://www.snmp-informant.com/that I load on the servers. Even the free basic client has a lot of great data. It basically takes WMI data and puts into SNMP so Orion can come along and grab. I have listed below all the data that I am now grabbing just by using the basic agent which is free.

SNMP Monitor-CPU
cpuDPCsQueuedPerSec
cpuInterruptsPerSec
cpuPercentDPCTime
cpuPercentPrivilegedTime
cpuPercentProcessorTime
cpuPercentUserTime
SNMP Monitor-Disk
lDiskAvgDiskQueueLength
lDiskAvgDiskReadQueueLength
lDiskAvgDiskSecPerTransfer
lDiskAvgDiskWriteQueueLength
lDiskDiskTransfersPerSec
lDiskPercentDiskTime
lDiskPercentIdleTime
lDiskSplitIOPerSec
SNMP Monitor-Memory
memoryPageFaultsPerSec
memoryPagesPerSec

We have also purchased several advanced agents which is $50 a server. And we are grabbing this data also.

cpuPercentInterruptTime
serverWorkItemShortages
swqAvailableThreads
swqQueueLength
systemProcessorQueueLength
tcpV4ConnectionsEstablished

Using this agent I have been able to grab about any performance metric that I would ever need from one of our servers. The company also has agents for Exchange, SQL etc. but I have not purchased any of these. However if you wanted to grab very specific data about SQL for example you could but the SNMP Informant agent for SQL.

By using SNMP informant this has allowed me to even more turn Orion into a one stop show for all of our monitoring needs. Using custom MIB I can even grab temperature status and power supply status for our Dell servers if needed.

chris.lapoint

Thanks to everyone that has posted thus far! This feedback will be extremely helpful as we plan our server monitoring and application monitoring enhancements for 2008. If you aren't already included in the new APM module beta, you'll be receiving an invite immediately ;-)

Thanks again!

FormerMember

1. If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?
We currently are trying to replace our multiple network/server monitoring systems with one system. A single focal point to get all information on our network whether it is a network switch or a server. If it is on the network then we want to manage it with one system. Currently we have HP Openview, HP OVIS, Whats Up Gold, MRTG and Kiwi - far to many systems to manage.

2. What other capabilities would you like to see us add around server monitoring and why?
Deeper server capabilities - especially WMI. We use WMI to collect serial numbers off of our servers - a simple thing but it sure beats going to each server to gather the information. We also sue WMI to tap into Exchange (and other performance monitors) for graphing and trending. The more information the better.
I also did not see a way to be able to unmonitor a device during a specific period. For example if a node needs to be monitored during business hours but not after hours I do not see where to schedule this in Orion. Don't get me wrong (because I am a newbie) but you provide a ton of capabilities into the latest version - it is just that I have run into a few "'ism's" already.

3.Why would you want to monitor the performance of applications like Exchange, SQL, etc in Orion?
Currently we are using multiple systems to do this. I want to be able to go to a single system to get a view of any device on the network. It is important to our system administrators to be able to trust the EMS system and be able to get a quick view and then more details as needed to trend issues. So far I have seen valuable functionality in Orion that the others don't provide.

4. When and how do applications fall into your realm of responsibility as a network engineer?
I currently maintain all of the above systems and have been trying to break free from the multiple system syndrome for awhile. When I saw Orion I was ecstatic that I might have a single tool to accomplish my job. One of the many factors why we use Whats Up Gold for so long is mainly due to the way the company was tuned into the needs of their customers and the forum network they setup and encouraged where users shared ideas, code and helped each other out was what made it such a successful tool at our location.

Keep up the good work

FormerMember

I also did not see a way to be able to unmonitor a device during a specific period. F

Go to Node Details for the target node and click the "Unmanage" button.

RSimmons

Hello,

We have just started using Orion to monitor our internal network (of about 40 servers) and external network (of about 10 x 384 port switches). We tested the Application Monitor plugin during our Evaluation period, but in the end chose not to purchase it due to the fact that it didnt provide the Layer 7 application monitoring.

What would we be looking for? HTTP, SQL, Exchange, SMTP monitoring should cover our needs - and the ability to create alerts based on service failures.

If these were covered then we would seriously consider expanding our SolarWinds usage and even look at providing the service to our customers (if its good enough).

Cheers,

Dakaix.

FormerMember

Nope - that only suspends polling until a set time whcih means that I would have to do this every day.

I need to be able to schedule non-polling for a device between say 6 PM until 7 AM.

FormerMember

I need to be able to schedule non-polling for a device between say 6 PM until 7 AM.

We're working on that functionality now.

chris.schear

If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?

If nothing else, to prove/disprove server related issues are not "network issues". Our network and server groups constantly butt heads on what a core cause for a problem is. "Must be a network issue", "must be a server issue", etc. The ability to have key insight into network and server health, statistics, and status is critical to end finger pointing. Also, we monitor lots of server's EtherChannel connections to determine load, load balancing issues, and general network performance concerns. Ultimately, servers communicate across the network and a holistic view into all end nodes is needed to effectively troubleshoot.

When and how do applications fall into your realm of responsibility as a network engineer?

There are some server-based applications which are inherently "network-related" in nature. Some examples would be TACACS+/RADIUS services, LDAP, DNS, DHCP, etc.

dmdc

If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?
What other capabilities would you like to see us add around server monitoring and why?
Why would you want to monitor the performance of applications like Exchange, SQL, etc in Orion?
When and how do applications fall into your realm of responsibility as a network engineer?

1) In my facility, there is not distinction between "network engineer" and "sysadmin." Everyone is expected to know how to do both. That said, we are currently monitoring about 1300 interfaces and majority of them are on servers. Since a large portion of our environment is related to renderfarming, it's very important for us to track both CPU and RAM utilization over time. The people in operations love the ability to create a report in Orion showing CPU load then export it to Excel and determine how efficient our renderfarms are. The SysAdmins use Orion to get a quick snapshot of RAM utilization to make sure new software we write internally isn't causing any memory leaks. When renderboxes lock up, we use the Alerts to notify the Syadmins of a failure so that the system can be rebooted. Really, Orion has changed the way the Engineering department works here.

2) I would love it if Solarwinds worked better "out of the box" with all the flavors of Linux. For example, if you install a default RedHat 4 Enterprise on a server, Solarwinds will be enable to monitor the network interface traffic. It involves painfully trying new versions of net-snmp which often leads us to change the Redhat kernel and then that breaks some of our software. We have similar problems with SuSE Enterprise.

3) In a renderfarming environment, it would be nice to track CPU load of the specific render application instead of the cumlative CPU usage of the server. For example, sometimes there is a CPU spike as a result of a CIFS transfer. It would be nice to just view the specific CPU data on an application level.

4) I guess I answered this in part 1.

JustinY

1. If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?

a. I would consider myself a network admin. I administer the network, and server infrastructure where I work. Paying for two separate programs to monitor the network, and servers isn't practical when we can use one single application that will do both.

2. What other capabilities would you like to see us add around server monitoring and why?

a. Ability to monitor system health of Active Directory , DFSR, Exchange 2007 Standby Continuous Replication, SQL 2005 database mirroring, port/service monitoring, and better performance counters. I would like to be able to see the health of our AD replication, and state of each domain controller. It would be nice to see the health of each of each DFSR file server in each replication group to see they are active, replicating, and healthy, Same for Exchange 2007 Standby Continuous Replication , and SQL 2005 database mirroring. Would beat having to run manual health checking tools, and would be nice to be alerted if health of those systems falls below metrics. Like if a domain controller fails, or fails for replicate for any reason.

3. Why would you want to monitor the performance of applications like Exchange, SQL, etc in Orion?

a. Orion is our primary network monitoring tool. Network as a whole not just connectivity. Being able to view the health of our network is nice as it allows our staff to see where problems might be from a single location.

4. When and how do applications fall into your realm of responsibility as a network engineer?

a. If anything fails its our responsibility.

njoylif

If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?
    In our environment, everyone is enjoying the capabilities and features Orion provides us in the network realm. It would be really handy to have a module that handled the server/application side of the house as in-depth as for networking...whether this burdon belongs with the same entity within SolarWinds or not is for you to decide. But I think it would be well received!
What other capabilities would you like to see us add around server monitoring and why?
   Perhaps some smart functionality for the more common services - NTP - do I pull an accurate time or am I even sync'd.     FTP, does the login work, does data transfer work.     HTTP, do a get and make sure the process is not hung, or even allow custom requirements so I can hit a DB on the backside and verify that is working - what good is a web front to the DB if th)DB is not puttin' out!
Why would you want to monitor the performance of applications like Exchange, SQL, etc in Orion?
See answer to first ?
When and how do applications fall into your realm of responsibility as a network engineer?
don't except for servers I'm running. See answer to first ?.

Thanks for taking feedback. I have to compliment SW for their excellence in this area of taking suggestings and working them in best as possible. Of course some things won't make business sense, but you're very good about putting in things that do!

patwill66

If you consider yourself to be network engineer (analyst, admin, consultant), why do you monitor servers in Orion today?
What other capabilities would you like to see us add around server monitoring and why?
Why would you want to monitor the performance of applications like Exchange, SQL, etc in Orion?
When and how do applications fall into your realm of responsibility as a network engineer?

1. I monitor servers with Orion because Orion is great at capturing and storing data for historical purposes. We are not a huge company and we dont have a huge network. We have under 10 people in IT and less than 200 network devices. Having a product set for network and another for server doesnt really make a whole lot of sense in such a samll environment. However, we are forced to use other free products that are out there (nagios, Cacti, Zenoss, Hyperic) because Orion cant monitor the critical services we need to get alert on. This has been by far the most disappointing aspect of the Orion software. If it wasnt for me liking the product so much, I sure the company would have dumped it by now because it cant do what we need. And this brings me to #2.

2. For monitoring, we need to be able to monitor the processes, services, ports, WMI, SQL and Exchange. Using nagios today, I am able to run a script that logs into my secure web portal and checks for specific text to be returned. If it isnt successful, it triggers an alert. I am able to query DNS and make sure I am getting the proper results backs. I am able to look at the WMI information for my Cisco CallManagers and get detailed information on call volume. Without being able to monitor some basic stuff on the server (service, process, port, WMI) I will always have to use two products, unless my company chooses to move to a more advanced product and get rid of Orion. I understand that an agent has to be loaded on the server to get WMI information, which I am fine with, but at least it will give people the option to monitor WMI if they like.

3. I would like to monitor the performance of Exchange and SQL in Orion because I would like Orion to be my central local for monitoring and graphing. It sucks have multiple products to monitor different things because not everyone is able to view the stats from these products. Currently myself and the other network engineer are the only two that can work with the Linux based monitoring tools so our Windows server guys need to also work with us to get updates and changes made.

4. Being a network engineer, I am resposible for network monitoring. Everything that touches the network I monitor. Again, not being a large enterpise network with a dedicated NMS guy, someone has to own monitoring and that would be me. So I would like to have Orion be as robust as possible because it makes my job easeier.