Integrating Prometheus and InfluxDB
summary
InfluxDB now accepts Prometheus’ remote_read and remote_write endpoints so scraped samples can be replicated into InfluxDB for long term storage, preserving metric name and tags and handling Prometheus-style queries. New execution engine and query language separate compute from storage to enable federated queries across multiple Prometheus and InfluxDB servers, pushdown of aggregations, and stateless query workers. Telegraph and Capacitor already scrape Prometheus targets, and work on Apache Arrow aims to speed bulk interchange; prototypes are in branches and 1.5 contains early plumbing.
AI-written summary · about
chapters (31)
- 0:00Introduction
- 1:30Data Model
- 2:30Multiple Data Types
- 2:56Query Language
- 3:22Prometheus Remote Read Write
- 3:41Why integrate
- 3:57Longterm storage
- 4:15Prometheus federated queries
- 4:28Prometheus push
- 4:48Stateless Prometheus
- 5:21Remote ReadWrite Prometheus
- 6:08Prometheus Queries
- 6:26InfluxDB Queries
- 7:16Ghetto Hack
- 8:00The Problem
- 8:38New Query Language
- 9:19Storage
- 9:43Stateless
- 10:01Scale
- 10:15Isolation
- 11:44Prometheus
- 13:33Parser
- 13:50Query Languages
- 15:23Prometheus Ecosystem
- 15:51Prometheus Tools
- 16:13Open Metrics
- 16:42Future Work
- 17:04Prometheus Connectors
- 17:49InputOutput Plugins
- 18:29Output Plugins
- 19:31Apache Aero
related talks
28mIntegrating Long-Term Storage with Prometheus [A]
33mIntro: Prometheus
31mEvolving Prometheus for More Use Cases
01h 03mWhat’s new in the Prometheus ecosystem?
42mPrometheus Intro and Deep Dive
36mPrometheus - Intro, Deep Dive, And Open Q+A
44mPrometheus in 2021 and beyond
01h 19mIntro + Deep Dive: Prometheus