Since running Wachahost.com, we have been trying to find a reliable website monitoring service that will send us SMS direct to my cell when our servers go down for whatever reason. So far it has been a mighty struggle to find a service that is both reliable, has the features we require (Not asking much here), and doesn’t cost an arm and a leg.
It may seem obvious in our case that when we are hosting sometimes 100+ websites on one server that we have monitors going around the clock making sure we are alerted whenever something goes wrong. But in many cases alerts can be used for just your everyday website when time is money. The biggest example would be if you were running a PPC/PPV campaign, and your website or host goes down. Often there is nothing you can do but ride out the downtime. But you can atleast pause your campaigns and not waste money pushing traffic to a downed websited.
When we first started looking for monitor services, we knew of only 2 features we required. Our website should be pinged every minute (Not every 5, not every 30, but every single minute of every day), it should send us SMS alerts at a moderate price. Seems simple enough. A number of services have a low limit of pinging every 5 minutes, this really isn’t good enough for us. We want to know immediately when our site is down. Ontop of that, our original plan was to provide uptime reports to our customers every month. Being upfront about how our servers were performing was going to be a big selling point of the business. As it turned out, and as I will explain, it didn’t quite work out that way.
We finally found a company that ticked all the right boxes. Monitis.com.
The biggest bonus about Monitis.com, is that you pay as you go. You don’t get locked into plans or given more monitors then you need. You basically build your plan as you go which was great for us when we were just starting out. Obviously we only had the one server at the beginning, but quickly grew and needed to be able to increase our plan as time went on.
So far so good. We set up the monitors to our servers, set up the alerts to go straight to our cellphones, and for the first month everything seemed OK.
Then, the downtime.
And no, it wasn’t our server going down but theirs. They actually have 3 datacenters spread over the United States, but that doesn’t help our uptime reports. You see each day I get sent an email with our uptime stats for the day. Almost every single day, one of their datacenters had failed. For example look at this image below :

I have coded this a bit to make it easier to understand. Just note that there is two rows per location as this is monitoring for TWO servers. We are only focusing on the first server. The top row shows that I had downtime that day, it is the cumulative stats for all 3 datacenters (Which follow) combined. Inside the star is how many failed loads my website gave. Basically suggesting my server was down. However as we go down, we check the first datacenter (Green box), hmm, seems this datacenter doesn’t think my server went down at all. We check the second datacenter, this datacenter thinks I have been down. And the last datacenter (Red Box), doesn’t think I was down either.
It doesn’t take a genius to work out that there was some issue in that middle datacenter today, and it is reporting that my server was down when infact it was not. It doesn’t worry me too much, but it goes back on one of our starting principles. How can I show the end of month uptime reports, when all the numbers are messed up and saying I had downtime when infact I had not? At one point we purchased a new shared server and it maybe sat with no more than a handful of customers on it, half of them not doing anything anyway. And it still said that server had gone down, despite the fact that the uptime on the server (Both with Apache and the server in general), said it had not gone down at all.
In anycase, there goes my idea to show uptime reports to our customers.

I didn’t want to cry over spilt milk, and other than cranky DC’s, the service was OK from Monitis. O wait, no it wasn’t.
It all started one night when I got a SMS at 3am that the server was down. I sleepwalked from my bed to my computer, and checked what was up. Nothing it seemed. The server had absolutely nothing wrong with it… When I checked Monitis, it said that my server was down for just one check, across all 3 datacenters. But it came up on the next check (1 minute later). Maybe the load had spiked on the server and it was unable to reach it at that exact point. Nope not that either. cPanel keeps load statistics for 15 minute and it was perfectly fine.
We had now moved on from just one Datacenter giving false downtime, to all 3 doing it. The SMS credits were only a few cents per text. So I didn’t worry about it too much. Other than having to get up at 3am and waking up my girlfriend in the process (There was rage, O the rage!), no harm done right? Well this started to get more and more frequent. It got to a point where I would get a text about the server being down, and there was a 50/50 chance it was a bogus report. Once again we are talking across all servers, even dormant ones so it was not like we were having random load spikes.
This all came to a head one night where I have never been so pissed off in my life.
There I am watching Mythbusters on a hot Thursday evening. And my phone goes off that a server is down. I check the server, seems OK. I go check the Monitis report, and for some reason the downtime is not even shown in the graph. It is showing 100% uptime. Even stranger, every time the server goes down I get a SMS ASWELL as an email. No email this time around. So I just got a SMS alert, with Monitis showing 100% uptime, the server showing 100% uptime, and no email to back it up.
Whatever. Mythbusters is calling.
30 minutes later. Same thing. And this goes on for several hours, me getting texts for a wide range of servers. None of which are showing any signs of downtime on monitis’s own reports, nor on mine. Monitis live support was non existant, they didn’t reply to our emails. I had to actually turn off all monitors so I could finally get some sleep. That is terrible. For a company to have to turn off their critical alert system, just to get some sleep. After waking in the morning, I turned on the monitors and everything was back to normal. However still showing zero downtime on the monitis backend, despite the inundation of sms messages on my phone.
Now we started to run into an issue. All these messages aren’t free. We had to pay every time something like this went on. And Monitis was never available on livechat for explanations. Admittedly I could have probably called them. And my Timezone probably wasn’t in check with theirs. But it was still frustrating as hell.
Things started to go really wrong. The usual, Get SMS but no email started to happen. Other times we got emails with no SMS. Sometimes the techs for Wachahost.com got texts but not me. Today my partner in the business said he got several emails about servers being down (I got none), and that every single downtime email was followed by a email stating that the server was back up… 1 minute later.
This has gone on long enough. Today we made the switch to Pingdom.com. They have a much smoother interface (Not Flash), an iPhone app, and so far, no false downtimes. I will write another post about my experiences with them at a later date, but for now, my advice is to stay away from Monitis.com