Hacker News new | ask | show | jobs
by Technophilis 2145 days ago
I am curious why you're transitioning off from Datadog? A monitoring solution based on Grafana and Prometheus tends to have a high initial setup cost plus on-going maintenance since it's self-hosted and lacks some of the features that Datadog offers out of the box (alert customization for instance).
1 comments

$$$

Custom metrics on Datadog are the main reason. We are a game company so we have all the metrics from the client games going through our server, which then send them to Datadog as custom metrics. This lets us monitor our client games in real time. When we started with Datadog many years ago it was fine, but then about 1.5 years ago they changed their billing to only allow 100 custom metrics per host. We were using about 2000 custom metrics per host. We had to aggregate the custom metrics down to an almost non-useful level to stay within the 100 custom metrics per host.

Also the $15 per host starts to add up if you want to run many smaller hosts rather than a handful of large ones.

I've been on many calls with my Datadog account rep and they refuse to budge on this. I would stay with them in a heartbeat if it was reasonably priced for what we want to do.

Yeah, I've heard custom metrics as the breaking point for a lot of DataDog users, TBH, so this doesn't surprise me.