alloy

grafanaApache-2.0活跃维护

基于 OpenTelemetry Collector 的可编程遥测数据管道,用声明式配置把指标、日志和链路统一采集、处理并转发到 Grafana、Prometheus、Loki 等后端。

OpenTelemetry Collector distribution with programmable pipelines

v1.21.3发布于 2026-10-28 18:32 UTC3,683star +5 / 7 天772fork同步于 2026-11-10 23:41 UTC

怎么上手

由 README 提炼,执行前请确认
README 未给出安装命令,请参考官方文档 https://grafana.com/docs/alloy/latest/get-started/install/
硬件要求按 README 说明

README 未说明

下载:源码 zip·GitHub Releases(安装包)本站不镜像安装包

项目说明

由 README 自动翻译提炼

Grafana 出品的 OpenTelemetry Collector 发行版,用可编程管道统一采集指标、日志、链路和性能剖析数据

基于 OpenTelemetry Collector 的可编程遥测数据管道,用声明式配置把指标、日志和链路统一采集、处理并转发到 Grafana、Prometheus、Loki 等后端。

核心能力

  • 用基于表达式的配置语法编写可编程可观测性管道
  • 兼容 OpenTelemetry Collector 的数十个组件
  • 对接 Prometheus、Loki、Pyroscope 等多种遥测生态
  • 原生支持 Kubernetes 资源,无需额外 Operator
  • 通过模块共享管道,并可组建集群自动分配负载

容易踩的坑

README 未给出安装命令,需按官方文档选择对应平台的安装方式;配置使用 Alloy 自有表达式语法,与原生 OTel Collector YAML 不通用

内容由模型根据 README 生成,可能不完整或不准确,以官方文档为准。看原文 README

最近发版

5 个版本v1.21.3
  1. v1.22.0-rc.0pre-release2026-11-05 · 2026-11-05 21:52 UTC
    查看更新说明

    Release Candidate 0 for v1.22.0

    This is a release candidate and should be used for testing purposes only.

    ⚠️ This is a pre-release. Do not use in production.

    Changes

    See the release PR #7436 for the full changelog.

    Upgrading

    Read the release notes for specific instructions on upgrading from older versions:

    Installation

    Refer to our installation guide for how to install Grafana Alloy.

  2. v1.21.3最新正式版2026-10-28 · 2026-10-28 18:32 UTC

    1.21.3 (2026-10-28)

    Bug Fixes 🐛

    • beyla.ebpf: Update Beyla to v3.38.1 (#7570)
    • database_observability.postgres: Improve table registry lifecycle and unqualified lookup (#7551) (895aa64) (@cristiangreco)
    • deps: Update github.com/andybalholm/brotli to v1.2.7 (#7546) (c913882) (@thampiotr)
    • Hide HTTP proxy env vars in support bundles (#7560) (014cd6a) (@ptodev)
    • loki.source.cloudflare: Stop additional_fields from corrupting the extended and all field presets (#7557) (3091601) (@vjsairam)
    • loki.source.kubernetes,loki.source.podlogs: Prevent re-ingesting logs on re-tails (#7512) (8cda438) (@kalleep)
    • mimir.alerts.kubernetes: Pick up Secret changes without a restart (#7555) (031f4eb) (@Churi12)
    • mimir.rules.kubernetes: Fix non-recovering unhealthy state after a failed restart when Mimir is unreachable (#7554) (a60d95c) (@TheoBrigitte)
    • otelcol.exporter.otlphttp: Deprecate enabled in the cookies block (#7552) (0e8389a) (@kgeckhart)
    • otelcol.exporter.prometheus: Pass missing include_scope_labels to converter (#7558) (52e7e14) (@mizuochik)
    • otelcol.processor.tail_sampling: Pass invert_match through for boolean_attribute policies (#7547) (5d7b5ea) (@kgeckhart)
    • otelcol: Report full build version including pre-release suffix in service.version (#7548) (401bf58) (@thampiotr)
    • prometheus.operator: Initialize controller-runtime logging (#7556) (ff0f144) (@mvanhorn)
    • remotecfg: Restore last success metric in notModified response (#7549) (df3d2fd) (@tpaschalis)

    Upgrading

    Read the release notes for specific instructions on upgrading from older versions:

    Installation

    Refer to our installation guide for how to install Grafana Alloy.

  3. v1.21.22026-10-22 · 2026-10-22 23:47 UTC
    查看更新说明

    1.21.2 (2026-10-22)

    Bug Fixes 🐛

    • Migrate config from type string to type secret [backport] (#7499) (6848963) (@blewis12)

    Upgrading

    Read the release notes for specific instructions on upgrading from older versions:

    Installation

    Refer to our installation guide for how to install Grafana Alloy.

  4. v1.21.12026-10-21 · 2026-10-21 20:49 UTC
    查看更新说明

    1.21.1 (2026-10-21)

    Bug Fixes 🐛

    • faro.receiver: Enforce download_from_origins on redirects and sourcemap URLs [backport] (#7481) (530886e) (@kgeckhart)
    • prometheus.exporter.kafka: Avoid panic on empty consumer group member assignment [backport] (#7441) (9b6b80a) (@ptodev)
    • Update go version to 1.27.3 [backport] (#7472) (88cfe69) (@kalleep)

    Upgrading

    Read the release notes for specific instructions on upgrading from older versions:

    Installation

    Refer to our installation guide for how to install Grafana Alloy.

  5. v1.21.02026-10-19 · 2026-10-19 11:24 UTC
    查看更新说明

    1.21.0 (2026-10-16)

    ⚠ BREAKING CHANGES

    • otelcol.receiver.kafka: otelcol.receiver.kafka and otelcol.exporter.kafka metadata > retry > max_retries now defaults to 20 (was 3) and is now actually applied. Upstream's Kafka client silently ignored it since the franz-go migration. otelcol.exporter.kafka now also applies backoff. (#7366)
    • otelcol.connector.signaltometrics: otelcol.connector.signaltometrics error_mode now defaults to ignore (was propagate). A failing OTTL expression now skips the record and logs the error instead of dropping the whole payload. Set error_mode = "propagate" to restore the previous behavior. (#7366)
    • loki.write: loki.write configurations that set wal { enabled = true } now fail to load unless Alloy runs with --stability.level=experimental. To keep using the WAL, add --stability.level=experimental to the alloy run command. To keep the default stability level, remove the wal block or set enabled = false. (#7257)
    • loki.write: Remove tenant label from metrics (#7232)

    Features 🌟

    • Add CAP_DAC_OVERRIDE in the allowed Linux capabilities list (#7303) (dddc5d8) (@grcevski)
    • database_observability.postgres: Enable stat_statements by default (#7224) (044d3a4) (@matthewnolf)
    • database_observability.sql_server: Add health_check collector (#7205) (c1d90e0) (@rgeyer)
    • database_observability.sql_server: Add support for clustering mode (#7208) (90100b3) (@matthewnolf)
    • database_observability.sql_server: Support monitoring multiple database blocks (#7207) (bbf24e2) (@matthewnolf)
    • database_observability: Add engine label for explain_plan log digests (mysql/postgres) (#6944) (551c3ad) (@rgeyer)
    • database_observability: Add relabel blocks to database_instance (#7369) (320c17e) (@matthewnolf)
    • helm: Add resizePolicy to alloy and config-reloader containers (#7274) (9983d74) (@younsl)
    • loki.source.cloudflare: Add api_url argument (#7353) (befe940) (@thampiotr)
    • otelcol.connector.signaltometrics: error_mode now defaults to ignore (was propagate) (#7366) (1e353a6) (@ptodev)
    • otelcol.encoding.jsonlog: Add the parse_ints argument (#7377) (42fe191) (@ptodev)
    • otelcol.exporter.loadbalancing: Add the randomness and `attrib

用户点评

登录用户的真实使用记录,可按版本和硬件筛选
共 0 条

    还没有人点评。第一个说出真实体验的人最有价值。

    README 原文

    未翻译

    Grafana Alloy is an open source OpenTelemetry Collector distribution with built-in Prometheus pipelines and support for metrics, logs, traces, and profiles.

    What can Alloy do?

    • Programmable pipelines: Use a rich expression-based syntax for configuring powerful observability pipelines.

    • OpenTelemetry Collector Distribution: Alloy is a distribution of OpenTelemetry Collector and supports dozens of its components, alongside new components that make use of Alloy's programmable pipelines.

    • Big tent: Alloy embraces Grafana's "big tent" philosophy, where Alloy can be used with other vendors or open source databases. It has components to perfectly integrate with multiple telemetry ecosystems:

    • Kubernetes-native: Use components to interact with native and custom Kubernetes resources; no need to learn how to use a separate Kubernetes operator.

    • Shareable pipelines: Use modules to share your pipelines with the world.

    • Automatic workload distribution: Configure Alloy instances to form a cluster for automatic workload distribution.

    • Centralized configuration support: Alloy supports retrieving its configuration from a server for centralized configuration management.

    • Debugging utilities: Use the built-in UI for visualizing and debugging pipelines.

    Example

    otelcol.receiver.otlp "example" {
      grpc {
        endpoint = "127.0.0.1:4317"
      }
    
      output {
        metrics = [otelcol.processor.batch.example.input]
        logs    = [otelcol.processor.batch.example.input]
        traces  = [otelcol.processor.batch.example.input]
      }
    }
    
    otelcol.processor.batch "example" {
      output {
        metrics = [otelcol.exporter.otlp.default.input]
        logs    = [otelcol.exporter.otlp.default.input]
        traces  = [otelcol.exporter.otlp.default.input]
      }
    }
    
    otelcol.exporter.otlp "default" {
      client {
        endpoint = "my-otlp-grpc-server:4317"
      }
    }
    

    Getting started

    Check out our documentation to see:

    Release cadence

    A new minor release is planned every three weeks. Patch releases are published every one to two weeks.

    The release cadence is best-effort: if necessary, releases may be performed outside of this cadence, or a scheduled release date can be moved forwards or backwards.

    Minor releases published on cadence include updating dependencies for upstream OpenTelemetry Collector code if new versions are available. Minor releases published outside of the release cadence may not include these dependency updates.

    Patch and security releases may be published at any time.

    Community

    To engage with the Alloy community:

    Contributing

    Refer to our contributors guide to learn how to contribute.

    Thanks to all the people who have already contributed!