3 ms·

The key constraint here is that the author has no access to persistent disks and can only use object storage for persistence. Otherwise Thanos would be extreme

• zzyzxd 5y ago
  > Thanos would be extreme overkill for this number of metrics Data volume is just one thing. Thanos makes Prometheus stateless and easy to shard, all in a non-invasive approach that is solid, boring, and just works. The architecture works well in small scale systems. I even use it in a single node k8s cluster in my homelab, pays only about ~$1 a month for Backblaze B2 so I never worry about data retention or disk usage. > Single-node VictoriaMetrics can easily handle 1M metrics/sec Even if I have disk access, I would think twice before deploying a database and manage it myself when I don't have to. Besides the maintenance burden and potential scaling issues in the future, it may cost you more to use block storage like EBS than S3. Also, Prometheus memory usage overhead for remote write was wild[1], so, good luck with capacity planning and config tweaking. 1. https://prometheus.io/docs/practices/remote_write/ https://prometheus.io/docs/practices/remote_write/
• znpy 5y ago
  > Single-node VictoriaMetrics can easily handle 1M metrics/sec yeah sure but then what will you be posting on Linkedin with buzzwords? "I have installed a boring software on a single machine because it works, performs well enough and it's cheap" ? what are you going to say, "I take daily snapshots of the vm because 1-day rpo is fine for me"? How will you get them likes?? (this is an ironic comment, before anyone starts a flame)