How to deal with Space issues #863
-
Is your feature request related to a problem? Please describe. Yes, local disk is getting used up, of course I can add more space but would like additional information on how harvest deals with rotating data. aka is there a way for us to specify how much old data exists? 1m/3m/6m before it removes old data or doesn't it currently remove old data out there? Describe the solution you'd like Describe alternatives you've considered Additional context |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments
-
hi @darthVikes - not sure if you saw this link in the wiki that discuss how much disk space Prometheus uses. The out-of-the-box default for Prometheus is to retain data for 15d. Some folks are also using VictoriaMetrics to store and/or downsample older data. |
Beta Was this translation helpful? Give feedback.
-
I ran into this same problem early on, and opened a discussion on it. The metrics have to reside somewhere, so it is up to you where you want the metrics to live. Many of the resources I had looked into, recommended keeping only recent data on the Prometheus servers. One factor in this decision is how much data you have and how many prometheus servers there are. In our case, we decided to leave the most recent 15 days of metrics on the Prometheus server, and then use Thanos to offload older metrics to object storage. Thanos also performs the downsampling and compaction that allows us to store long-term data efficiently. It was a lot more work to set this all up, but now that it is running it is easy enough to maintain and it works acceptably well. HTH! |
Beta Was this translation helpful? Give feedback.
hi @darthVikes - not sure if you saw this link in the wiki that discuss how much disk space Prometheus uses.
The out-of-the-box default for Prometheus is to retain data for 15d. Some folks are also using VictoriaMetrics to store and/or downsample older data.