Monitoring Services Are Poised for a Shake Up
2011/09/06 Leave a comment
Server monitoring and website monitoring services cost too much and are overly complicated.
Over the past several months we have built and launched NodePing’s site and server monitoring service. Part of that process involved looking at the other companies in this market niche, and finding the opportunities for offering a service that fills a gap in what is currently being provided to customers. What we have found has confirmed our original reasons for starting NodePing.
There are a lot of companies offering site and server monitoring services. However, our experience as consumers of these services was that it was hard to find a provider that did what we needed at a reasonable price, and I think our experience is probably typical. Where’s the disconnect? We wanted a service that would allow us to watch twenty to thirty sites and services for a reasonable price. It is easy for a small to medium business to get to a couple of dozen services needing monitoring. Most companies have at least one or two web sites that need to be available all the time for their customers. Many also have two or three web sites used internally for collaboration and sharing or publishing information to employees (Intranets). Throw in a DNS server or two, a mail service, an accounting system, a key router or two, and you are quickly into double digits on the number of services that need to be checked.
IT departments used to run software like Nagios for this type of thing, and that is still a good option in many cases. Nagios provides a wider set of checks than a typical SaaS monitoring service, there are lots of specialized plugins available, and it is not all that difficult to write custom plugins. If you need specialized checks, a system like Nagios is probably the best bet. On the other hand, while Nagios is free software, running it is not free. It requires a server to run on. Typically you want monitoring to run on separate infrastructure from your normal servers, which often means leasing a server or using a VPS service. Doing this inexpensively typically runs $50-100 a month, and involves a non-trivial amount of technical expertise and work to setup, tune, and maintain. That’s not a huge amount of money, but it is not free.
External providers offer similar services. The majority of companies need HTTP, SMTP, and PING checks. These are the primary checks provided by the bulk of the monitoring as a service industry. These types of services don’t cost much to run. With today’s opportunities to build and deploy cloud based services in cost effective ways that scale well, the cost of these types of services should be fairly low. That’s not currently the case.
A quick search turns up a lot of companies offering these services. Many of them offer “free” or inexpensive services. “Free” monitoring is typically provided for one to five URLs, often with fifteen to thirty minute intervals. That is basically useless. If it is ok for a service to be down for 30 minutes without getting a notification, you probably don’t need monitoring. In my opinion, a price “plan” isn’t a serious offer unless they offer the service in intervals of five minutes or less at that price. Getting beyond that unhelpful “Free” level, many providers start charging by the URL or address you want to monitor. One company prominently advertises checks starting at $1, but again that’s one URL in thirty minute intervals, and it costs $11 for that URL check if you want to do check it every minute. Paying per check or per URL quickly gets expensive. It is not uncommon to find special price calculators on the sites of this kind of provider, which is itself a hint that the pricing is too complicated. At these prices, a fairly typical small to mid-sized company could easily find themselves spending hundreds of dollars a month on monitoring.
There are more competitive options out there. These companies typically cost $40-$60 for a reasonable number of addresses and services. These prices probably save you money compared to running monitoring yourself using something like Nagios. Plus, you don’t need to deal with setting up and maintaining the software. That’s a pretty good deal.
However, it still doesn’t need to cost that much. With modern hosting and technology, the cost per check and even per customer to run these types of services is very low. In fact, just about the only cost of running a service like this that is attributable to an individual account is the credit card processing. All the rest of the costs scale, and are spread in ways that actually decrease per account as you scale up. Unless they are just running very inefficient systems, the total overhead for the companies charging $40-$60 per month (not to mention the ones costing hundreds) should be less than $4 per customer. Of course, the companies advertising “Free” services are also spending dollars a click to get those accounts, and that easily becomes the biggest expense. Meanwhile, allowing their customers to add additional checks or URLs to an account costs the provider pennies. Pricing based on adding checks or URLs is a model completely detached from the economics of running the service.
Experience in running IT departments and talking to system administrators tells us that there are a lot of services that should be checked if best practices were followed that aren’t getting checked. Many companies that use external providers check their company’s primary site, but when adding checks means adding overhead costs (or just the work load), secondary and internal sites don’t get checked. This means that there are millions of services that should be monitored that aren’t getting monitored at all. Companies are just reacting to complaints when something goes down.
To us, this smelled like opportunity. It is not simple to set up a solid monitoring service. However, once the technology, infrastucture and processes are in place, it is a service that scales. The margin stays fairly stable even if you let customers use it as much as they need. This calls for a flat rate model.
Our biggest problem is that we have entered a market that is saturated by misinformation. Buyers assume that this type of service costs at least $40 for a reasonable level of monitoring, and often lots more. They expect to see low entry prices that don’t really meet anybody’s needs, followed by much higher prices for the real service. This becomes a marketing challenge. When shopping for these services, NodePing’s price of a flat $10 for monitoring sounds like one of the entry point bait ads. We say “$10 to monitor up to 1000 services in 1 minute intervals” and people ask “Yes, but what do we really get, and what’s it cost if we actually need to do real world monitoring?”
NodePing’s services really cost $10 a month. Period. There are no add-ons, no “X is available at additional cost”. We set 1000 services as the maximum because we don’t want to monitor IBM’s network (no offense to IBM intended). Our target is small to medium sized businesses, and we want them to monitor everything they want to monitor for one reasonable price. If this model works, maybe others will also move to flat rates. That’s great. We’d be happy to help make the monitoring world make more sense and be more cost effective for businesses. We think we have a solid technology stack and a great service, and we can do quite well even if other providers compete with us directly on price. Until then, there are few if any major providers that really provide the services that our customers need anywhere close to our price.
Monitoring services cost too much and are too complicated. We think this market is set for a change, similar to how the cloud has impacted other technology services. This shift will be a significant benefit to small and medium sized companies that need these services, and it is a fantastic opportunity to providers poised to provide the services the customers need at truly competitive scale and rates. NodePing has positioned itself to provide the services businesses need at a fantastic, flat-rate price.