- 已经在用 OTel Collector 做埋点、但后端还挂在商业 SaaS 上：SigNoz 直接吃 OTLP，埋点代码不用改，换的是 exporter 指向和账单
- 排查跨服务延迟问题：日志、指标、trace 在同一个数据模型里关联，从 APM 面板下钻到具体 span 不用在三个工具之间对时间戳
- 需要保留遥测数据所有权（合规、数据出境、成本不可控）：自托管在 K8s 里，数据落在自己的列式库，代价是运维这套存储
- 中小团队想同时要 APM + 日志 + 告警而不想拼 Grafana + Loki + Tempo 三件套：省掉跨组件关联的配置，但换来的是单点依赖
signoz
SigNoz活跃维护 有专家点评基于 OpenTelemetry 的开源可观测性平台，把日志、指标和链路追踪集中到一处，提供 APM、分布式追踪和基础设施监控，帮团队快速定位线上问题。
SigNoz is an open-source, OpenTelemetry-native observability platform for your team and their AI agents. Get logs, metrics, and traces in one tool with features like APM, distributed tracing, log management, infra monitoring, etc. Combined with SigNoz MCP and a native AI teammate (in SigNoz Cloud) it helps you build more resilient apps.
v0.149.0发布于 2026-11-04 06:55 UTC32,574star +30 / 7 天2,590fork同步于 2026-11-10 23:23 UTC
专家点评
上手难度：进阶更新于 2026-11-11OpenTelemetry 原生的自托管可观测后端，用来替换 Datadog、New Relic 这类 SaaS APM 的采集与查询层
适用场景
踩坑
- 自托管没有 Noz AI 助手，那是 Cloud 专属；如果评估时被 AI 排查能力吸引，先确认这条不在社区版里
- README 没给可直接复制的最小启动命令，Docker、K8s、Linux 三条路径要分别读文档；想 clone 完就跑起来会卡在部署方式选择上
- issue 里反复出现的是周边集成诉求（Grafana API、ClickHouse 存储后端、Coolify 部署、OIDC 社区版支持），说明核心链路之外的企业集成和认证能力是长期缺口，上生产前先确认你的 SSO 和现有 Grafana 面板能不能接
- 平台本身不采集数据，必须自己部署 OTel Collector 或 Prometheus 接入，空跑起来是个没有数据的空壳
同类对比
Grafana 栈（Loki + Tempo + Mimir）在已有 Grafana 生态、愿意自己拼装关联逻辑时更灵活；纯 trace 场景下 Jaeger 更轻。要托管省心就直接用商业 SaaS，自托管的运维成本是真实开销。
自托管请参考官方文档：https://signoz.io/docs/install/self-host/ 硬件要求按 README 说明
README 未说明具体 CPU、内存或显存要求，仅说明可自托管在 Docker、Kubernetes 或 Linux 上运行。
项目说明
由 README 自动翻译提炼开源 OpenTelemetry 原生可观测平台，把日志、指标、链路追踪和告警集中到一处
基于 OpenTelemetry 的开源可观测性平台，把日志、指标和链路追踪集中到一处，提供 APM、分布式追踪和基础设施监控，帮团队快速定位线上问题。
核心能力
- 用 OpenTelemetry 原生采集，一次埋点保留遥测数据所有权
- 关联日志、指标、链路与异常，从服务图表直接下钻排查
- 内置 APM 面板展示延迟、错误率、吞吐与 Apdex
- 支持分布式追踪火焰图、瀑布图与 Trace Funnels 漏斗分析
- 基于单一列式数据库承载高基数高吞吐观测数据
容易踩的坑
README 未给出可直接复制的最小安装命令，需按 Docker/Kubernetes/Linux 对应文档分别部署；Noz AI 助手仅在 SigNoz Cloud 提供，自托管社区版没有；数据需自行用 OpenTelemetry 或 Prometheus 接入，平台本身不自动采集。
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
- v0.149.0最新正式版2026-11-04 · 2026-11-04 06:55 UTC
[!WARNING] We are moving log bodies to native JSON storage. From this release, logs are dual-ingested, which requires additional collector CPU and memory, ClickHouse CPU, and log disk space. Plan for this capacity before upgrading. Learn more here https://signoz.io/guides/json-logs-migration and track it here https://github.com/SigNoz/signoz/issues/13376
🚀 Features
- feat(stats): report AI observability usage from the stats endpoint (#13359) @nityanandagohain
- feat: add events for span detail open and trace view (#13358) @tewarig
- feat(authz): rule view FGA (#13251) @nikhilmantri0902
- feat(dashboards): add plugin schema for scatter plot panel (#13030) @ahrefabhi
- feat(roles): pick the grant key in the telemetry selector wizard (#13226) @vikrantgupta25
- feat(authz): downtime schedule FGA (#13250) @nikhilmantri0902
- feat(traces-explorer): fetch all span fields in the list view (#13275) @aks07
- feat(authz): support service and environment keys in telemetry selectors (#13036) @vikrantgupta25
- feat(traces): return instrumentation scope in waterfall, flamegraph and list view (#13280) @krsoninikhil
- feat(storybook): query builder component stories (#13325) @ahrefabhi
- feat(aiobservability): read the messages the Langfuse SDK writes (#13331) @nityanandagohain
🐛 Bug Fixes
- fix(telemetrystore): return invalid input with a fix for three ClickHouse SQL mistakes (#13328) @srikanthccv
- fix(query-builder): legacy having casing and non-positive limits (#13315) @ahrefabhi
- fix(ai-assistant): open explorer chips in the view the query implies (#11733) @makeavish
- fix(notification-channels): scroll channel forms inside configuration and pin actions to the bottom (#13382) @ashu8912
- fix(alerts): papercut fixes for routing policies and planned downtime tables (#13320) @ashu8912
- fix(querier): stop zero-start OOM and reject malformed time windows (#13246) @tushar-signoz
- fix(channels): reset previous type's template values on type switch (#13374) @ashu8912
- fix(traces-qb): synthesized attribute keys follow the JSON column evolution (#13307) @krsoninikhil
- fix(dashboards): label unstacked bar layout as overlap (#13369) @ahrefabhi
- fix(saved-views): rename old camelCase trace select fields (#13326) @krsoninikhil
- fix(saved-views): remove loading effect from authz checks (#13287) @H4ad
- fix: noz mention email (#13319) @tewarig
🧰 Maintenance
- chore(deps): bump axios to 1.21.0 (#13339) @H4ad
- chore(frontend): remove unused dependencies (#13312) @H4ad
- chore(query-builder): add data-testids for e2e selectors (#13311) @ahrefabhi
- test: fix apt-get not found error in integration tests (#13367) @namanv3
- test(e2e): retry clicks swallowed by dnd-kit and loading antd buttons (#13401) @H4ad
- test(query-builder): stop the QueryV2 Delete menu test timing out in CI (#13394) @AshwinBhatkal
- docs(go): require descriptive receiver names (#13391) @vikrantgupta25
- refactor: using foundry commands for aws and azure integration (#13270) @swapnil-signoz
- ci(e2e): report the e2e check on merge queue runs (#13381) @AshwinBhatkal
- test(e2e): rebuild the suite on a driver kit (#13189) @H4ad
- perf: divide cache buckets for a query into cells in a grid (#13333) @namanv3
- refactor(saved-views): key by source page instead of data source (#13296) @tewarig
- refactor(dashboard): group panel type picker into components and hooks (#13170) @ahrefabhi
- refactor(react-router): route every react-router call through a src/lib/router facade (#13173) @H4ad
Full Changelog: https://github.com/SigNoz/signoz/compare/v0.148.1...v0.149.0
- v0.148.12026-10-29 · 2026-10-29 17:25 UTC
查看更新说明
What's Changed
- docs(e2e): point playwright agents and contributing guide at drivers (#13309) @ahrefabhi
- test(integrations): stop GCP account settings drawer tests timing out under load (#13310) @H4ad
🚀 Features
- feat(alerts): show routing policies and planned downtime as tables (#13259) @ashu8912
- feat(alerts): add configuration sidebar for channels, routing policies, and planned downtime (#13230) @ashu8912
- feat(authz): route policy FGA (#13249) @nikhilmantri0902
- feat(onboarding): add Hono, Caddy, PM2, and GCP one-click datasources (#13291) @crazyuploader
- feat(tracedetail): read waterfall/flamegraph attributes from JSON column (#13274) @krsoninikhil
🐛 Bug Fixes
- fix(sentry): drop errors from third-party and injected scripts (#13304) @H4ad
- fix(notification-channels): match the relaxed channel checks and pick webhook auth with a radio (#13276) @AshwinBhatkal
- refactor(query-builder): drop stale bar-panel data-source override (#13218) @ahrefabhi
🧰 Maintenance
- chore: preemptively update rules v2 schema's api spec properties (#13306) @namanv3
- chore(saved-views): retire the old views pages, v1 client and the flag (#13258) @aks07
- v0.148.02026-10-28 · 2026-10-28 06:44 UTC
查看更新说明
🚀 Features
- feat(alerts): move anomaly alerts to the new alert experience (#13130) @AshwinBhatkal
- test(dashboards): e2e for panel and editor flows (#12410) @ahrefabhi
- feat(saved-views): saved views and quick filters VQA fixes (#13232) @aks07
- feat: llm o11y improvements (#13229) @tewarig
- feat(saved-views): analytics for saved views and the bottom strip (#13164) @aks07
- feat: adding gcp cloud integration agent mode changes (#13103) @swapnil-signoz
- feat: trace details thread (#13073) @tewarig
- feat(alerts): adopt server-side clone API in frontend (#13181) @ashu8912
- feat(saved-views): saved views and quick filters VQA fixes (#13232) @aks07
🐛 Bug Fixes
- fix(query-builder): explorer crash on unsupported query in url (#13257) @aks07
- fix(serviceaccount): delete role mappings when a service account is deleted (#13266) @vikrantgupta25
- fix: fix time series caching edge cases (#12813) @namanv3
- fix(settings): remove notification channels from settings nav (#13225) @ashu8912
- fix(querier): limit scalar metric series after reduceTo (#13083) @ahrefabhi
- fix: make related logs/traces link carry the relevant query from a rule (#13112) @namanv3
- fix: put non-positive formula heatmap values in the zero bucket (#13217) @namanv3
- fix(alertmanager): treat missing dotted labels as nil in downtime scope (#13118) @praneeth-signoz
🧰 Maintenance
- chore: add ai qb examples (#13282) @nityanandagohain
- chore(openapi): describe the API payload fields and add request examples (#13200) @srikanthccv
- chore(quick-filters): remove the old checkbox filter (#13262) @aks07
- chore(skills): add authz-configurator skill (#13261) @vikrantgupta25
- chore: bump cloud integration agent version (#13245) @swapnil-signoz
- chore(codeowners): add owner for saved views and the bottom strip (#13243) @aks07
- chore(claude): require page stories with frontend UI changes (#13231) @vikrantgupta25
- chore(telemetrystore): read JSON natively instead of as a string (#13009) @krsoninikhil
- test(dashboards): panel stories with play assertions for every kind (#13244) @ahrefabhi
- refactor(quick-filters): one quick filter source drives the saved list and the fields api signal (#12832) @aks07
- test(dashboards): e2e for viewing, sections, time range, variables and settings (#13264) @ahrefabhi
- test(dashboards): e2e smoke checks for every panel kind (#13273) @ahrefabhi
- test(dashboards): e2e for the dashboards list page (#13263) @ahrefabhi
- test(dashboards): e2e flow for the Time Series panel (#13233) @ahrefabhi
- test(dashboards): e2e drivers and spec builders (#12409) @ahrefabhi
- test(dashboards): expose the uPlot instance on the chart container (#13271) @ahrefabhi
- chore(deps): bump handlebars,brace-expansion,storybook (#13267) @H4ad
- refactor(workspaced-locked): avoid using images from signoz.io (#12880) @H4ad
- refactor(trace-details): minor refactor (#13067) @tewarig
- chore(telemetrystore): read JSON natively instead of as a string (#13009) @krsoninikhil
New Contributors
- @ashu8912 made their first contribution in https://github.com/SigNoz/signoz/pull/13181
Full Changelog: https://github.com/SigNoz/signoz/compare/v0.147.0...v0.148.0
- v0.147.02026-10-22 · 2026-10-22 06:42 UTC
查看更新说明
🚀 Features
- feat(notification-channels): add the telegram channel form (#13119) @AshwinBhatkal
- feat(authz): enable FGA for metrics explorer routes (#13150) @vikrantgupta25
- feat: enabled logs and added collection strategy for gcp integration (#13184) @swapnil-signoz
- feat(notification-channels): move the channels UI onto the v2 API with a redesigned list (#12976) @AshwinBhatkal
- feat(notification-channels): rebuild the channels list on the v2 API (#13128) @AshwinBhatkal
- feat(alerts): Support telegram channel for alerts (#13074) @praneeth-signoz
- feat(saved-views): turn on saved views and the bottom strip (#13125) @aks07
- feat(trace-details): waterfall tooltip redesign and AI span badges (#13066) @tewarig
- feat(logs): move the list row filter chips off v3 attribute keys (#13148) @aks07
- feat(saved-views): gate saved views on their permissions (#13027) @aks07
- feat(ruler): add clone alert rule API (#13093) @nikhilmantri0902
- feat(querier): read traces json numbers as toFloat64OrNull and index that (#13101) @krsoninikhil
- feat(promote): index materialized paths on their json sub-columns (#13054) @krsoninikhil
- feat(charts): mark partial time buckets with a dotted ring (#13165) @ahrefabhi
- feat(authz): enable FGA for infra monitoring routes (#13143) @vikrantgupta25
- feat: use allow list in getSpanDisplayData (#13124) @tewarig
- feat(tracedetail): format raw Anthropic and Gemini bodies, skip embeddings (#13141) @nityanandagohain
- feat(tracedetail): format OpenRouter answers and Vercel AI SDK prompts (#13139) @nityanandagohain
- feat(tracedetail): format LangChain serialised runs (#13137) @nityanandagohain
- feat(statsreporter): open the stats endpoint to every authenticated caller (#13152) @vikrantgupta25
- feat(tracedetail): format OpenAI Responses bodies (#13136) @nityanandagohain
- feat(tracedetail): format openllmetry legacy prompt and completion keys (#13135) @nityanandagohain
- feat(tracedetail): format OpenInference flattened messages (#13133) @nityanandagohain
- feat(tracedetail): format gen_ai system instructions and finish reasons (#13131) @nityanandagohain
- feat(components): adopt @signozhq/ui 0.2.0 across the app (#12971) @manika-signoz
- feat: return formatted gen ai message in waterfall (#13121) @nityanandagohain
- feat(query-builder): field placeholders, default row limit and synced fields (#13113) @ahrefabhi
- feat: update ai attribute mapping (#13120) @nityanandagohain
- feat: thread rows by the gen_ai gate, formatted tool spans (#13115) @nityanandagohain
- feat(trace-details): move the span filter and span actions off v3 query APIs (#13134) @aks07
- feat(query-builder): move group-by and order-by key suggestions to the v5 fields API (#13127) @aks07
- feat(saved-views): hide the old saved views ui behind the flag (#13028) @aks07
- feat(saved-views): all views panel in the quick filters sidebar (#13022) @aks07
🐛 Bug Fixes
- fix(notification-channels): apply the review follow-ups on the channel screens (#13214) @AshwinBhatkal
- fix(metricreduction): type the no-kept-labels preview count so the driver can scan it (#13221) @vikrantgupta25
- fix(exceptions): keep the filter pills when running the query (#13204) @aks07
- fix(explorer): stop reporting dropped order by to sentry (#13153) @aks07
- fix(tracedetail): keep litellm tool calls and AI SDK JSON answers, mark redacted media (#13166) @nityanandagohain
- fix(onboarding): link docs by final URL and dedupe swift data source (#13161) @crazyuploader
- fix(alerts): drop vulnerable v1 rule history endpoints (#12163) @therealpandey
- fix: hide create alert in case of ai query (#13122) @tewarig
- fix(onboarding): link docs by final URL and dedupe swift data source (#13161) @crazyuploader
🧰 Maintenance
- chore: drop sloglint key-naming-case (#13220) @namanv3
- chore(e2e): address tinypool audit (#13205) @H4ad
- chore(notification-channels): remove the v1 channel client and the old list (#
- v0.146.02026-10-14 · 2026-10-14 06:42 UTC
查看更新说明
🚀 Features
- feat(logparsingpipeline): order user pipelines around signozlogsnormalizer (#13106) @tushar-signoz
- feat(tracedetail): normalise gen_ai messages in thread spans (#13003) @nityanandagohain
- feat: add support for AI queries in Dashboard (#12833) @namanv3
- feat(dashboard): support variables in panel title (#12988) @ahrefabhi
- feat(onboarding): add Bun/Elysia, SNMP, Vercel Eve, Microsoft Agent Framework, Android datasources (#13100) @crazyuploader
- feat(promote): authorize promoted paths per signal (#13029) @krsoninikhil
- feat(promote): support indexes on trace attributes and remove prefix requirement in paths (#13021) @krsoninikhil
- feat(promote): refactor path promotion API to support traces as well (#12839) @krsoninikhil
- feat(infra-monitoring): filter k8s lists by pod status, node readiness and container status (#12946) @AshwinBhatkal
- feat(trace-details): read header metadata from the trace summary API (#13065) @tewarig
- feat(alert-channel-integrations): integration tests using wiremock for google chat, incident.io, jira and jsm ops- #12435 (#12985) @praneeth-signoz
- feat(saved-views): views header in the quick filters sidebar (#13011) @aks07
- feat(infra-monitoring): add the overview tab to the kubernetes side drawer (#12948) @AshwinBhatkal
- feat(dashboard): redesign panel editor config pane (#13010) @ahrefabhi
- feat(dashboard): new-panel drawer with section placement and type browser (#12992) @ahrefabhi
- feat: trace detail thread endpoint (#12999) @nityanandagohain
- feat: add trace summary endpoint (#12911) @nityanandagohain
- feat(quick-filters): saved views slots in the sidebar layout (#12994) @aks07
- feat: add heatmap support in dashboard schema (#12844) @namanv3
- feat(ruler): add saved views CRUD APIs for the alerts list page (#12814) @nikhilmantri0902
- feat(promote): support indexes on trace attributes and remove prefix requirement in paths (#13021) @krsoninikhil
🐛 Bug Fixes
- fix(dashboard): stop panel type drawers flashing on open and close (#13114) @ahrefabhi
- fix(querier): apply the limit to scalar trace operator queries (#13091) @ahrefabhi
- fix(tests): use promoted_paths API in create_json_index fixture (#13105) @tushar-signoz
- fix(explorer): preserve order by in list view for explorers (#13085) @aks07
- fix(home): crash due to race condition on load motion (#12681) @H4ad
- fix(dashboard): heatmap legend search and two-row cap (#13082) @ahrefabhi
- fix(new-select): reveal truncated values and contain dropdown scroll (#13063) @AshwinBhatkal
- fix(logs): log details header shifts when switching to the Context tab (#13070) @aks07
- fix(promql): serve transpiled series without their synthetic name (#13031) @namanv3
🧰 Maintenance
- chore(claude): run py-fmt and py-lint after python changes (#13116) @tushar-signoz
- chore: delete unwanted models from pricing db (#13079) @nityanandagohain
- chore(storybook): add more stories for specific pages/components (#13001) @H4ad
- chore(storybook): replace story-shots scripts with sbshot (#13075) @H4ad
- chore(storybook): add new-panel picker and config pane stories (#13058) @ahrefabhi
- chore: replace banner in llm with callout (#13037) @tewarig
- chore(deps): bump urllib3 from 2.7.0 to 2.8.0 in /tests (#13032) @dependabot
- chore(deps): bump grpc to v1.83.2 and opentelemetry-go to v1.45.0 (#13047) @srikanthccv
- chore(deps): bump google.golang.org/grpc from 1.83.1 to 1.83.2 in /scripts/promqltestcorpus (#12804) @dependabot
- ci(e2e): shard playwright across 4 runners and install browsers in background (#13077) @H4ad
Full Changelog: https://github.com/SigNoz/signoz/compare/v0.145.0...v0.146.0
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。
SigNoz is an open-source observability platform built on OpenTelemetry. We’re building an enterprise-grade alternative to fragmented monitoring stacks, with logs, metrics, traces, alerts, and dashboards in one place.
Choose how to run SigNoz
SigNoz Cloud (Recommended)
Fully managed SigNoz with a 30-day free trial, no credit card required, usage-based pricing that starts at $49, and regional data hosting.
Enterprise
Enterprise Cloud, BYOC, or Enterprise Self-Hosted with compliance, support, custom retention, RBAC, ingestion controls, data residency, and region selection.
Community
Free open-source SigNoz that runs in your own infrastructure. Deploy with Docker, Kubernetes, or Linux and keep full control of your data plane.
What can you monitor?
SigNoz helps teams debug production issues faster by connecting logs, metrics, traces, alerts, dashboards, exceptions, and agent-native workflows in one place.
APM Overview
Monitor service latency, error rate, throughput, Apdex, top endpoints, database calls, and external calls.
Learn more: APM documentation
Log Management
Ingest, search, aggregate, and correlate logs with traces and metrics using a visual query builder.
Learn more: Log management documentation
Metrics and Dashboards
Build dashboards for application, infrastructure, and custom metrics using Query Builder, PromQL, or ClickHouse SQL.
Learn more: Metrics documentation
Infrastructure Monitoring
Monitor Kubernetes clusters, pods, nodes, workloads, and host-level CPU, memory, disk, network, logs, and traces.
Learn more: Infrastructure monitoring documentation
LLM and AI Observability
Trace LLM apps, RAG pipelines, prompts, tool calls, tokens, latency, and costs alongside application and infrastructure telemetry.
Learn more: LLM observability documentation
Agent-Native Observability and MCP
Use the SigNoz MCP server to bring telemetry into coding agents, or use Noz inside SigNoz to investigate incidents, tune alerts, and build dashboards with production context. Noz is available only on SigNoz Cloud.
Learn more: SigNoz MCP server docs · Agent skills docs
Distributed Tracing
Follow requests across services with flamegraphs, waterfalls, span events, filters, and trace analytics.
Learn more: Distributed tracing documentation
Trace Funnels
Create funnels from traces to understand request-flow drop-offs, failed transitions, and systemic workflow issues.
Learn more: Trace funnels documentation
Also monitor: exceptions, alerts, external APIs, and integrations for OpenTelemetry, Prometheus, Kubernetes, cloud providers, language SDKs, application frameworks, databases, and LLM tools.
Why teams use SigNoz
- OpenTelemetry-native Instrument once with open standards and keep ownership of your telemetry.
- Correlated signals Move from service charts to traces, logs, infra metrics, and exceptions without switching tools.
- Single columnar database Built for high-cardinality, high-volume observability workloads.
- Predictable pricing No per-host pricing, no user-seat pricing, and no special pricing for custom metrics.
- Enterprise ready SOC 2 Type II and HIPAA compliance, RBAC, ingestion controls, custom retention, support, BYOC, and self-hosting.
Getting started
Start on Cloud
Create a managed SigNoz workspace and get your first dashboard without running observability infrastructure.
Self-host SigNoz
Run SigNoz in your own infrastructure with Foundry, Docker, Kubernetes, or Linux.
Foundry · Docker · Kubernetes · Linux
Send data
Instrument applications and infrastructure with OpenTelemetry, Prometheus, language SDKs, and integrations.
Instrumentation · Integrations
Comparisons to familiar tools
SigNoz is often adopted by teams moving from a stack of single-purpose tools or commercial platforms with unpredictable pricing.
Prometheus Good if you just need metrics. SigNoz keeps metrics, logs, traces, dashboards, and alerts together so teams can debug with correlated context.
Jaeger Jaeger only does distributed tracing. SigNoz adds metrics, logs, trace analytics, dashboards, alerts, exceptions, and trace-to-log workflows.
Elastic SigNoz uses columnar database for efficient observability analytics and high-cardinality log workloads, with 50% lower resource requirement compared to Elastic during ingestion. Check the detailed study.
Loki In the linked benchmark, SigNoz indexed all keys in the test setup, while Loki hit max stream errors when more labels were added. Check the detailed study.
Contributing
We ❤️ contributions big or small. Please read CONTRIBUTING.md to get started with making contributions to SigNoz.
Not sure how to get started? Just ping us on #contributing in our slack community.
As always, thanks to our amazing contributors!
langfuse
langfuse·TypeScript
开源 LLM 应用可观测与评估平台，能追踪调用链路、跑评测实验，帮团队定位问题并持续优化提示词和模型效果。
LLM 可观测+4LLM 可观测进阶
37266+1424.2kv4.78.02026-11-10 23:15 UTC
skywalking
apache·Java
分布式系统的应用性能监控平台，支持链路追踪、指标采集和日志分析，帮你在微服务架构下快速定位性能瓶颈。
应用性能监控+4LLM 可观测硬核
24999+56.6kv11.1.02026-11-10 12:30 UTC
jaeger
jaegertracing·Go
CNCF 出品的分布式追踪平台，采集服务调用链路并做可视化分析，帮你在微服务和 LLM 应用里定位延迟与故障根因。
分布式追踪+4LLM 可观测进阶
23395+93.2kv2.23.02026-11-10 17:10 UTC
openobserve
openobserve·TypeScript
开源可观测性平台，统一收集日志、指标、链路、RUM 和会话回放，单二进制部署，存储成本比 Datadog、Splunk、Elasticsearch 低约 140 倍。
可观测性平台+4LLM 可观测进阶
22980+991.2kv1.2.0-rc12026-11-10 21:00 UTC