To give you an idea of my 3rd party monitoring experience, we started with BinaryCanary then to NodePing then added Pingdom for our public uptime reporting (something NodePing is lacking in) but switched from Pingdom to StatusCake because of Pingdom's monitor limit for their free plan and because StatusCake had a really nice feature for embedded reporting which allowed us to make a custom report for our status page. To clarify, we've paid for both BinaryCanary and NodePing while using the free plans for Pingdom and StatusCake.
We have no plans of ditching NodePing, the monitoring and alerts have been spot on every time and 1000 1-minute monitors for $15/month is an unbelievable value. Their public reports are amazing but are way to detailed for end users to find useful which is why we use another service for our status page.
Enter StatusCake...
StatusCake looks very impressive and we made an awesome public report from their embedded report, unfortunately the false alarms generated are too frequent forcing us to delete the monitors and remove them from our status page.
Today I woke up to 180+ StatusCake alerts that all of our data center's were offline but NodePing, Pingdom (still monitoring our Tampa location), 1 custom internal monitor, and 1 custom external monitor reported no packet loss all day. I am assuming these false alarms are due to the fact we only have the free plan so it's only using 1 monitoring server but the inaccuracy has caused us a lot of headaches because we had StatusCake as our public uptime report and clients were complaining about downtime they never experienced.
We've considered purchasing a paid plan, but we're hesitant to invest money into something that might not work. I've been waiting to see if the false alarms would improve but today they were so bad I just deleted all of the monitors out of anger. Since we've started using them a few weeks ago, we'd get a handful false alarms a day for a few minutes at a time. At one point last week it showed our networks with over 1 hour downtime in a single day and clients, who did not experience any downtime, were requesting SLA credit for the "downtime". We provided them reports from NodePing and Pingdom showing the monitoring was inaccurate which some accepted but I can only imagine how many took the StatusCake reports and shared them with their contacts or will use them in reviews they write about us.
Originally I thought that this was a complete fluke with out networks and had ignored it for a while hoping things would clear itself up, but then I went onto #lowendbox a few nights ago and found others reporting the same issues (I thought one of them even had a paid plan but I cannot confirm this).
What prompted me to write this review was that a lot of people on IRC and here have commented on how great StatusCake is and I really hope my experience is not the norm. I saw the new thread about the "Website Hosting Uptime Report" and noticed another LEB provider listed on there (one that signed up the same day as me and using it for the same reasons I was) and I happen to have a VPS with this provider that I use for monitoring and network troubleshooting, this provider shows a 98.9759% uptime out of 10,253 checks meaning they failed about 106 checks which my records do not reflect.
While this review may seem completely negative it was not my intention, I am posting it since I noticed @SCDaniel is active on here and would like for him to be aware of the situation I have experience and if others have a similar experience he can work on resolving it.
StatusCake is a really clean service and the presentation is the best I've seen from a monitoring company (NodePing could take some notes in that department). I really hope that my experience is not the norm which is why I welcome any positive reviews for this service so I can feel comfortable about picking up a paid service to continue using their awesome embedded reporting feature.
I would like to take the opportunity to make a suggestion that may improve the quality of service for free account users...
BinaryCanary's billing structure is based on the number monitors but this also counts each monitoring server as one monitor. So for example if a server is being monitored for SSH (port 22) and HTTP (port 80), that is 2 monitors. If you have 2 monitoring servers checking those 2 services then it's a total of 4 monitors. If you choose to only have 1 monitoring server monitoring those 2 services, they will automatically use a 2nd monitoring server as a double check if the first monitoring server reports either service down.
I know the above paragraph might be confusing but my suggestion is that if a monitoring server reports downtime then have a 2nd server double check that report to see if that monitoring server itself is accurate, if that monitoring server returns a false alarm then possible remove it from the monitoring pool until it can be manually checked for accuracy. I know this would require some complex code depending on the current backend setup, but it's just a suggestion for the future since your free plans are the gateway to getting people to upgrade to a paid plan.