The value proposition of all cloud services is that you get instant technical capability without building your own infrastructure. I see cloud NMS services like LogicMonitor and Spiceworks as a good deal for small organizations without their own IT people. But for all the reasons you give, the model doesn't scale very well. For network professionals, the value of self-managed internal monitoring infrastructure far outweighs the temporary ease and low cost of cloud monitoring. In particular, commercial monitoring offerings, such as Intermapper, PRTG, and SolarWinds, are extremely cost effective for business network operations. Their cost is easily justifiable, especially if you have a busy staff. Yes, you can get many of the commercial tool capabilities in open source projects such as OpenNMS and Cacti. But as you note, they can be a pain to configure, and if your labor is worth anything, the commercial options are usually a better deal. One exception I've found recently is Mikrotik's The Dude, which is free, but not FOSS. It's fully graphical, is straightforward to install and configure. It has a client/server architecture like Intermapper, but doesn't run natively on as many platforms (Windows only; other OSes must use emulation). Although it works with any SNMP device, it has special support for Mikrotik, since Mikrotik devised it. To recap, I think cloud monitoring is pointless for managing inside networks for any organization having a reasonably capable IT staff. On Jan 28, 2015, at 10:06 AM, Jay Hennigan <jay@west.net> wrote:
I know that this topic has been kicking around for at least a decade, but wanted to get current opinions of other network operators. Most of us have explored Nagios, MRTG, and several front-ends for MRTG.
We are looking into a new player in the space called Logicmonitor. They have a very functional and easy to navigate front end and configuration tool, and I very much like the look-and-feel of their product.
What I don't like is that they only offer it as a cloud-based service. Internal probes tie in to a "collector" which we maintain. The collector then phones home over the Internet to their hosted service periodically and they remotely analyze the data and generate alerts, plot graphs, etc.
From a technical standpoint this adds more points of failure in series, will cause missed alerts if their cloud-based service goes down (who is guarding the guards?) will cause false alarms if their service is still up but can't reach the collector, and doesn't give us a full view under the hood.
Of course their sales guys are giving us "Our time and energy is dedicated to reliability" and "professionally managed multi-carrier highly secure data centers" language to encourage the warm fuzzies.
From a scalability standpoint we incur ever-increasing recurring costs as we grow and add monitored devices and services.
What's the collective opinion here? Is anyone using them or a similar service? Are there non-cloud-based alternatives that are relatively easy to set up and manage? We've explored Zabbix, Nagios, MRTG and its various wrappers, and Intermapper. Anything else new on the horizon that has a GUI front-end that is configurable without a lot of scripting experience, etc.?
We would love to buy something that works for us and pay a reasonable price for it, but I'm not particularly interested in the equivalent of renting a time-share in order to monitor our networks.
-- Jay Hennigan - CCIE #7880 - Network Engineering - jay@impulse.net Impulse Internet Service - http://www.impulse.net/ Your local telephone and internet company - 805 884-6323 - WB6RDV