Reconnoiter is a new tool in the world of monitoring. It is not only a trending tool but also an alerting/fault detection one. IMHO, I reckon that it's in the trending capacities that Reconnoiter has a very huge potential.
One of the premises Recoinnoter is built upon is that RRDTool large installations are very inneficient regarding I/O usage and I think this is RRDTool major problem.
One of the things that can difficult the change from the traditional RRDTool frontends to Reconnoiter is, obviously, the cost of change and the learning curve.
So, any of you has experience with Reconnoiter? How's the learning curve? Was it difficult to move from RRDTool frontend applications (Cacti, Munin, Ganglia) to Reconnoiter? I'm looking forward to read your opinions.
The really neat thing about Reconnoiter is that graphs are effectively pointers to the metrics they contain. There's no archaic concept of rrd database files for each graph, just a JSON object describing the graph.
There is a learning curve with installation and setup of Reconnoiter, which may be a drawback. Alternatively, you should check out Circonus (http://circonus.com/), which is a hosted service built on Reconnoiter, adding in fault detection and notifications. They have videos up on the site.