Observability Awesome

If you can't measure it, you can't manage it. If you can't measure it, you can't improve it.

-- by Peter Drucker

When a measure becomes a target, it ceases to be a good measure.

-- by The Risk of Goodhart's Law

• o11y - Observability
• getsentry/sentry笔记Sentrygetsentry/sentry · BSL, Python · error tracking and performance monitoring platform打开：/notes/service/observability/tracing/sentry
• Uptime
  • louislam/uptime-kumaGitHublouislam/uptime-kumalouislam/uptime-kuma · MIT, JS, Vue · 监控网站 uptime · https://demo.uptime.kuma.pet/笔记：uptime-kuma
    • MIT, JS, Vue
    • 监控网站 uptime
    • demo.uptime.kuma.pet
    • uptime-kuma笔记uptime-kumalouislam/uptime-kuma · MIT, JS, Vue · 监控网站 uptime · https://demo.uptime.kuma.pet/打开：/notes/service/observability/uptime-kuma
  • upptime/upptimeGitHubupptime/upptime
• Alerting
  • ccfos/nightingaleGitHubccfos/nightingale
    • Apache-2.0, Go
  • opsre/WatchAlertGitHubopsre/WatchAlert
    • MIT, Go
    • Alerting for Prometheus
  • prometheus/alertmanagerGitHubprometheus/alertmanagerprometheus/alertmanager · https://prometheus.io/docs/alerting/latest/alertmanager/ · /api/v2 · https://petstore.swagger.io/?url=https://raw.githubusercontent.com/prometheus/aler...笔记：alertmanager
• Storage
  • trickstercache/tricksterGitHubtrickstercache/trickster
    • Apache-2.0, Go
    • HTTP Reverse Proxy Cache and Time Series Dashboard Accelerator
• openobserve笔记OpenObservezinclabs/openobserve · Apache-2.0, Rust · 原 ZincObserve · 从 ZincSearch 演化而来 · Log & Trace & Metrics · 存储 Parquet + Datafusion打开：/notes/service/observability/openobserve
  • zincobserve -> openobserve
  • Apache-2.0, Rust
  • 从 ZincSearch 演化而来
  • Log & Trace & Metrics
• 通用时间序列数据库/time-series笔记Database Awesome| name | stand for | · | Relational DBMS | 关系型数据库 | · | Key-value stores | KV 存储 | · | Document stores | 文档存储 | · | Time Series DBMS | 时序数据库 |章节：时序数据库打开：/notes/db/awesome
• opentelemetry.io/ecosystem/vendors
• observIQ/bindplane-opGitHubobservIQ/bindplane-op
• hyperdxio/hyperdxGitHubhyperdxio/hyperdx
  • MIT, TS
  • Clickhouse 存储
  • session replays, logs, metrics, traces, errors
• clickvisual/clickvisualGitHubclickvisual/clickvisual
  • UI for Clickhouse
• highlight/highlightGitHubhighlight/highlightApache-2.0+EE, TS, Go · error monitoring, session replay, logging, trace · https://github.com/highlight/highlight/blob/main/docker/compose.yml笔记：highlight
• keephq/keepGitHubkeephq/keepkeephq/keep · MIT+EE, Python, Typescript · FastAPI, SQLite, PostgreSQL, Redis, Elasticsearch, Pusher, Socket.IO · React, NextJS, React Chrono, Chart.js, Recharts, TailwindCSS笔记：Keep
  • MIT+EE, Python, Typescript
  • AIOps and alert management platform
• prymitive/karmaGitHubprymitive/karma
  • Apache-2.0, Go, Typescript
  • Alert dashboard for Prometheus Alertmanager
• google/mtailGitHubgoogle/mtail
  • Apache-2.0, Go
  • 从应用程序日志中提取监控指标
• Logging / Trend / LLM Observability
  • openobserve/openobserveGitHubopenobserve/openobserve
    • 开源可观测性平台，覆盖 logs、metrics、traces、frontend monitoring、pipelines 和 LLM observability。
  • sansan0/TrendRadarGitHubsansan0/TrendRadar
    • AI 驱动的舆情与趋势监控工具，聚合多平台热点、RSS 和推送渠道。

Tracing

• teletrace/teletraceGitHubteletrace/teletrace
  • Apache-2.0, Go
• hypertrace/hypertraceGitHubhypertrace/hypertrace

Metrics

• Prometheus
• InfluxDB
• victoria-metrics笔记VictoriaMetricsVictoriaMetrics/VictoriaMetrics 是什么？ · Apache-2.0, Golang · 时序数据库 · 高性能、低成本、大规模 · 支持集群 · 默认推荐单节点 - 小于 百万/s 的指标速率推荐单节点版本 - 通过垂直扩容/增加 CPU 内存来提升性能打开：/notes/service/observability/metrics/victoria-metrics
  • 最好，最推荐
• m3db/m3笔记M3m3db/m3 · Apache 2.0, Go · Distributed TSDB · 集群依赖 ETCD · Why · Why Choose M3DB for Logz.io Prometheus-as-a-Service · vs. Thanos & Cortex打开：/notes/service/observability/metrics/m3
  • Apache-2.0, Go
  • Distributed TSDB, Aggregator and Query Engine, Prometheus Sidecar, Graphite Compatible, Metrics Platform
• 基于/fork Prometheus 的服务
  • promxy笔记promxyjacksontj/promxy · 聚合多个 prometheus 进行查询 · config.yaml · nocache=1 for VictoriaMetrics QueryParams · VictoriaMetrics优化: · 添加 查询参数避免缓存问题打开：/notes/service/observability/metrics/promxy
  • cortex笔记Cortexcortexproject/cortex · horizontally scalable, highly available, multi-tenant, long term storage for Prometheus. · 架构 · 提供远程写入接口打开：/notes/service/observability/metrics/cortex
  • thanos笔记Thanosthanos-io/thanos · Prometheus 长期存储 · 后端使用对象存储服务 · 多租户方案 · sidecar 配置 external label，使用多层 query · receive 接收多租户，但不推荐 · 适用于网络隔离场景，querier 无法直接访问 sidecar 时打开：/notes/service/observability/metrics/thanos
• jacksontj/promxy笔记promxyjacksontj/promxy · 聚合多个 prometheus 进行查询 · config.yaml · nocache=1 for VictoriaMetrics QueryParams · VictoriaMetrics优化: · 添加 查询参数避免缓存问题打开：/notes/service/observability/metrics/promxy
  • 聚合多个 prometheus 进行查询
  • config.yamlGitHub Filejacksontj/promxy/cmd/promxy/config.yaml
    • nocache=1 for VictoriaMetrics QueryParamsGitHub Filejacksontj/promxy/pkg/servergroup/config.go#L96-L99
• prom-migratorGitHub Filetimescale/promscale/cmd/prom-migrator
• grafana/mimirGitHubgrafana/mimir
  • AGPLv3, Go
  • 长期存储，使用 S3 存储
• 商业
  • Datadog - APM 和监控平台
  • New Relic - 应用性能监控
  • AWS CloudWatch - 云指标
• spinnaker/kayentaGitHubspinnaker/kayenta
  • Automated Canary Service
  • part of Spinnaker
• argoproj/argo-rolloutsGitHubargoproj/argo-rolloutsargoproj/argo-rollouts · Apache-2.0, Go · Rollout 可替代 Deployment · Blue-Green / 蓝绿发布 · Canary / 金丝雀发布 · 精细化流量权重分配 · 自动回滚与自动推广笔记：Argo Rollouts
  • 利用 Kubernetes 生态，声明式地管理发布过程
  • 渐进式交付（Progressive Delivery）

APM

• uptrace/uptraceGitHubuptrace/uptrace
  • BSL, Golang
  • 存储使用 ClickHouse
• SigNoz/signozGitHubSigNoz/signozSigNoz/signoz · MIT+EE, Golang, React · APM, ClickHouse · 应用信息存储在 SQLite #941 · dashboards layout, alert config · signoz/alertmanager笔记：Signoz
  • MIT, Golang
  • OpenTelemetry Collector, ClickHouse
  • signoz.io/docs/architecture
• opensearch-project/observabilityGitHubopensearch-project/observabilitylogs - 日志 · metrics - 指标 · session replays - 会话重放 · traces - 跟踪 · errors - 错误 · 目前业内共识 OTEL · 大多场景直接选择 OpenTelemetry 即可 · If you can't measure it, you can't manage it.笔记：Observability Awesome
• Graphite
• Javamelody
• Stagemonitor
• Scouter
• Elastic APM
• App Metrics
• Glowroot
• erda-project/erdaGitHuberda-project/erda
  • Apache-2.0, Go

国人开发维护

• sunface/datavGitHubsunface/datav
  • Apache-2.0, TS+Go
• dianping/catGitHubdianping/cat
  • Apache-2.0, Java
  • by 美团点评
• pinpoint-apm/pinpointGitHubpinpoint-apm/pinpoint
  • Apache-2.0, Java
  • HBase
• apache/skywalkingGitHubapache/skywalkingapache/skywalking · Apache-2.0, Java · Application Performance Monitoring System · OAP - Observability Analysis Platform笔记：SkyWalking
  • Apache-2.0, Java
  • ES, H2, MySQL, PostgreSQL
• deepflowio/deepflowGitHubdeepflowio/deepflow
  • Apache-2.0, Go, Rust

Profiling

1. 更加细致的指标监控 - 方法，进程级别
2. 更加细粒度 - 毫秒级
3. 主要用于程序优化

• grafana/pyroscopeGitHubgrafana/pyroscope
  • AGPLv3, Go
• parca-dev/parcaGitHubparca-dev/parca
  • Apache-2.0, Go
  • Continuous profiling for analysis of CPU and memory usage
• andrewrk/poopGitHubandrewrk/poop
  • MIT, Zig
  • Performance Optimizer Observation Platform
  • linux perf_event_open
• plasma-umass/cozGitHubplasma-umass/coz
  • BSD 2-Clause, C
  • Causal Profiling
• ZaMaZaN4iK/awesome-pgoGitHubZaMaZaN4iK/awesome-pgo