application monitoring devices
we're an ISV ready to release our SaaS application online to our end customers, and also are presently seeking an application monitoring remedy. Along with monitoring the common OS-level suspects (I/O, disk room, logs, CPU, RAM, exchanging, and so on ), we're additionally aiming to check, sharp and also record on inner application occasions, problems, and also counters (assume line up dimension for inner solution, or latency of a solution we're obtaining from a 3rd party using personalized APIs).
We're begun considering Nagios, Zenoss, and so on, yet figured out those do just low-level things, and also are presently considering MOM and also ManageEngine. Still, they are much from being a personalized application monitoring device.
So - do you have anything to recommend?
You need to consider Quest Foglight. It has the typical web server - based metrics yet additionally application monitoring for both J2EE,. Net, along with all the large application collections (PeopleSoft, E - Business Suite, etc). One attribute that could be a great suitable for you is the Foglight Experience Monitor. It records http demands off the cord (agentless) and also records on feedback times at the various layers of the technology pile. You can utilize this to check those 3rd parties you rely upon and also can also give this as a valued - included attribute to your consumers that can after that check the performance of your SaaS offering.
I make use of Nagios to check those really points, along with disk room and also load standards and also all that. See, things concerning Nagios is that it is simply a mistake monitoring system, and also it relies upon plugins to in fact examine anything. It takes place ahead with some wonderful default plugins for monitoring SNMP variables and also pings and also suchlike, yet you can check anything with it, as long as you can write a program to examine something and also return an OK/warning/critical/ unidentified sign.
The 'plugin' language is a little bit foolish, given that the 'plugin API' total up to regulate args, setting variables, an exit standing, and also what you publish on stdout. I have greater than a couple of plugins that are simply 20 - line ksh manuscripts.
StackHub does what you are asking : "monitor, sharp and also record on inner application occasions, problems, and also counters". It is a SaaS itself, so you do not need to install and also config anything cumbersome like Nagios/Zenoss/etc
The solution remains in beta now and also originally targets Java applications (as an example, Log4J/Logback), so very early responses rates.
A pair opportunities:
Most monitoring systems - from Nagios to Zenoss to HP OpenView - permit you to write personalized displays, which might be what you desire.
You can write a less complex display yet make it extra monitoring - system - agnostic by exporting it making use of the modification attributes of (linux) net - snmpd. These allow you export whatever arbitrary numbers or strings you desire out over snmp ; all you need to do is write the tiny manuscript to bring the numbers themselves from your application. This can be from log documents or a standing documents that your application draws up occasionally or whatever.
Perhaps your Software as a Service application can result SNMP details, which can be accumulated by any kind of variety of devices (Nagios, Munin, etc).
Sunlight Java System Messaging Server is an instance of an application that gives a great deal of data using SNMP. The SNMP Implementation area of their Admin Guide claims:
The energetic details concentrates on presently queued messages and also open network links (as an example, matters of queued messages, resource IP addresses of open network links), while the historic details gives collective total amounts (as an example, complete messages refined, complete incoming links).