As those of you that are active in our beta programs know, we're about to release a new add-on to Orion that does "Application Performance Monitoring". The goal here is to provide some monitoring capabilities within Orion that provides application status and performance metrics from the needs perspective of the network engineer.
One of the things that I'm constantly challenged with is users complaining that they can't get to an application (that application could be making a phone call, surfing to YouTube, accessing the intranet, Exchange, a database server - whatever) because "the network is down" or "the network is slow". This new add-on should, among other things, provide me with the ability to know first if there actually is a network problem that is affecting applications and/or if the application itself is actually down.
All that said, in many cases I'm finding that the network staff is also being tasked with maintaining servers and applications and therefore their application monitoring needs are deeper than what I first envisioned. To that end, we've extended the capabilities of this new module and we will be continuing to do so over the next several months.
I'm curious how many of you are tasked with managing or monitoring applications and for those of you that are responsible for this, what level of depth you're looking for...
Thanks,Josh
We mainly monitor applications in an effort to absolve the network.
Sometimes there is an application which is highly unstable and all teams are focused on stabilizing the app.
I will monitor the servers & network links involved with Orion to assist in troubleshooting the app.
I have a custom property called App & create custom views based on this, so that all elements can be viewed on a single page.
If and when the application is stable, I will remove the monitoring.
Our network team primarily monitor the network but when users complain about slow response on the network for an application we then start looking at the application and server to absolve the network from being the cause of the problem. It's not until we can prove one way of another that its the application\server can any changes be made to either by the server support team. We're in this ongoing two way fight always trying to prove its not the network that is at fault. So any tool that can help with this would be great. We currently use IPMonitor but would love to have it all in Orion.
Any application slowness is always pinned on the network first. So we always get the call first and have to prove it's not the network. The only issue with this is we really can't see what the user is experiencing unless we share a screen with them, etc. Josh..this is why I made that suggestion of a product (not an agent) that could be launched locally on the user's machine (ad-hoc) that we in support could configure to test certain things, and provide other details we need to look into the issue. The users are really not technical enough to get this information to us quickly becuase most of the slowness issues go away before we can look at it.
I dont manage servers, nor do I want to.. however I am monitoring application response time via macro's with products like citratest. Essential I have a few laptops in key segments in the network that run a macro I created to do transactions in our applications like retrieve client information or payment process.
I time each step from login to form submits and retrieves.. every product out there is junk and I love to see you guys come up with something that could integrate into Orion and eventually the netflow module.
Well find me a network engineer that doesn't need to monitor application response times and you'll find me a liar! :D
We have an ongoing fight between blame on network response and that from the applications. As most of you will no doubt experience users log calls to the our helpdesk blaming the network as being slow when it really is and when actually there response from the business application they are using is problematic. As a network team we then have to trawl the network in search of problems (or lack of) to then report into the application support teams that the network is or isn't at fault. In our situation (hosting company) this situation is even worse as the applications are supported by external companies or other internal group companies and are therefore not within the same reporting structure.
I for one would therefore like to see Solarwinds develop a tool to perform configurable user style tests of key applications the biggest being web based. Having a report showing timed stages for each stage of a path that included logging into a website, browsing to a defind page and then performing a set search on that page would allow my team to collate user experience against technical information available via SNMP (network usage, CPU usage, memory etc...).
Looking forward to seeing what you have planned.
Hi,
I've been using Solar winds and IP Switch's WhatsupGold for network monitoring over 80 network/server devices for performance and uptime plus automated responce 24x7. Between the two I have found Solar Winds to be a far superior, enhanced tool that any admin cant ignore. I would certainly suggest Solar Winds to any one starting out to monitor his critical systems and networks even if remotely.
Akash.Bhardwaj
I did like to bring your notice to all the members here- I have been using Solar Winds and IP Switch's WhatsupGold for network montoring. I have over 9 remote offices in Asia Pacific region to look after and over 80 critical enterprise servers and network devices.
Of the two I have found Solar Winds to be the better, enhanced tool.
Akash
Hi, it is possible that there are incidents that can be caused by the network in production although more often than not does it turn out to be a transaction based problem. being able to pin-point from a transaction not only where in the network a bottleneck or issue happens but also correlating transaction information with the application, network, database and operating systems has help minimize incidents in production for us.