Remote Read Meets Streaming
Prometheus Blog

Remote Read Meets Streaming


Summary

Prometheus 2.13.0 introduces a streamed version of the Remote Read API, addressing significant memory usage issues with the previous implementation. The new API utilizes a chunked, iterative approach—similar to how Prometheus processes queries internally—allowing clients to receive data series by series without buffering the entire response in memory. This improvement enables more efficient querying of Prometheus data, especially for large datasets and use cases like seamless upgrades and integration with long-term storage systems like Thanos.
Read the Original Article

This article originally appeared on Prometheus Blog.

Read Full Article on Original Site

Popular from Prometheus Blog

1
Interview with iAdvize
Interview with iAdvize

Brian Brazil • May 17, 2017 • 118 views

2
When (not) to use varbit chunks
When (not) to use varbit chunks

Björn “Beorn” Rabenstein • May 8, 2016 • 118 views

3
Custom service discovery with etcd
Custom service discovery with etcd

Fabian Reinartz • Aug 17, 2015 • 118 views

4
Interview with Scalefastr
Interview with Scalefastr

Brian Brazil • Feb 8, 2018 • 114 views

5
Interview with Canonical
Interview with Canonical

Brian Brazil • Nov 16, 2016 • 112 views