I'm really glad you've noticed this post and that librato will continue to improve based on the feedback.
I hope my post will NOT get across as "picking librato in the first place was a mistake", because we definitely don't regret starting with you guys. We were up and running quickly with metrics and alerting in no time and we got to learn how important it is to be able to filter metrics per host. We definitely didn't had the time to set up everything by ourselves and your platform + great support provided the right solution.
At the time, the alternative was to set up our own graphite and from my experience, internal tools that you don't have time to _really_ invest in, will break faster than you can think of. Having dealt with non-maintained, internally owned self managed graphite+syren stack before, I have 0 regrets :)
But as we grew and InfluxDB matured, we've done a bit of research, discovered cool features (e.g. tags) we can make great use of and later on decided to go with managed InfluxDB solution and the ecosystem it brings along.
In the post I also argue that librato is a good pick, if you're a small team with no ops people and need a metrics+alerting backend ASAP.
