Infrastructure and endpoint monitoring is the latest addition to the suite of Unistellar services.
Late last year we were fat, dumb, and happy using a service called Monitis for our server and website monitoring. I think I learned about them originally through my previous employer and I spent about $250 annually for 5 servers and 2 endpoints.
Suddenly and inexplicably our account was locked. I was advised that Monitis was sold to TeamViewer and that I should proceed there to continue my service. I could not find anything that resembled what I liked about Monitis through TeamViewer … and I was a little jaded because of the abrupt transition … so I started researching alternatives.
- Uptrends
- My favorite of the options, but only fit about 50% of my needs
- Interface was pleasant and intuitive, but they had an awkward demarcation between their infrastructure and endpoint dashboards. It was also incredibly slow to find performance counters for server agents like they never heard of caching
- Fred (the owner) was pleasant and available but when I brought up concerns I could tell he had no interest in improving what he already had or even getting feedback
- The way they handle endpoint testing is kinda fishy – you can’t select specific cities like you could with Monitis… and Paris, Mumbai, and Kansas City all gave me about the same RTT to a box that was in Kansas City!
- No way to embed infrastructure monitoring graphs on sites like NotifyPrice’s Status Page
- Costs twice as much as Monitis did – and you have to get a quote for the infrastructure side of things by phone or email
- Agent was lightweight and quick to install
- dotcom-monitor
- A close second … interface was good and they actually figured out the caching bit for making performance counter configuration fast
- Crazy expensive for 1 minute interval testing… everything they offer defaults to 4 hours which is absurdly unuseful
- Pingdom
- You have probably heard about SolarWinds in the news … this is their product for website monitoring
- If the above doesn’t make you cringe, the site reported 25 minutes of downtime before the first successful check
- Not only do you not know what cities you are getting tested from, but you also only get to pick test locations on a continent level
- CloudRadar
- A tie for second in my mind … they basically give you a limited list of Azure locations … and you can only pick one which has inherent propensity for false alarms
- Their server agent was a little heavier at just over 11 MB
- Datadog
- Probably overkill if all you need is infrastructure and endpoint monitoring
- Most expensive option on the list (only work with Enterprise level clients)
- Server agent was huge and took forever to install … I think that is as far as I got before moving on
- Uptime
- I cannot say with any certainty that they do anything beyond endpoint monitoring
So at the end of the day (looking at maybe $500 a year) I decided to write my own Windows Server agent. And with that success I launched an ingest API. And the natural progression was to add a UI around it on the web. After I got users, notifications, roles, some reporting, and maintenance windows fleshed out it was just a couple days to do the invoicing and payment methods. Overmonitor was born!
- Cloud-based
- Small and lightweight agent
- Fast configuration interface
- Completely customizable pricing
- Maintenance windows that work
- Geotargetting at the city level
- All graphs embeddable
- Push notifications for any alert
- Process monitors with rollup
- The use case here was NotifyPrice offline scrapers… 32 of them run all with the same process name and different command line arguments
- Trend analysis and capacity planning
- After I put the configurable low and high alert thresholds, it occured to me I could do a daily trend analysis using linear regression on the samples we had collected for a given metric
- The above coupled with a planning horizon will alert users about the possibility of running out of processing power, memory, or disk space in the months ahead
- Network configuration alerts
- Seemed like a “no brainer” but I don’t remember any of these vendors telling me when subnets or external IP address changed or are mixed (subnet) or duplicated (external IP)
- Distributed monitors
- Allows users with agents installed on servers to configure their participation in our growing network of test locations
