# harmonycloud/kindling

eBPF-based Cloud Native Monitoring Tool

Repository: https://github.com/harmonycloud/kindling
Canonical: https://ross.abutalabs.com/products/kindling
Homepage: http://kindling.harmonycloud.cn
Language: Go
License: Apache-2.0
License Family: permissive
Topics: kubernetes, monitoring, observability, profiling, ebpf
Last push: 2026-07-21T08:12:43+00:00

## Health v2 (maintenance only)
Score: 65/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 93, release rhythm 8, longevity 100
- inputs: {"age_days": 1710, "days_push": 43, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1062, forks 166 (observed 2026-08-28T04:03:26.227233+00:00)

## What it is
Kindling is an eBPF-based cloud native monitoring tool that runs as a DaemonSet in Kubernetes to capture syscalls and tracepoints from kernel to application code. It provides trace profiling (OnCPU/OffCPU thread-level flame graphs correlated with logs and network/file metrics), network flow visualization, and Prometheus/Grafana integration for root-cause triage.

## Use cases
- monitor kubernetes network flows like tcp retransmit dns and throughput
- profile distributed traces at thread level with oncpu and offcpu events
- find root cause of slow rpc calls in production
- correlate logs and flame graphs with trace execution timeframes
- visualize service maps with latency error rate and traffic golden metrics
- enrich prometheus metrics with ebpf-based full stack monitoring
- diagnose thread starvation and lock contention issues

## When to choose
- you run workloads on linux kubernetes and need kernel-level observability without code instrumentation
- you want to correlate traces with cpu offcpu events logs and network syscalls
- you need network-level rpc analysis similar to chrome devtools network panel for backend services
- you already use prometheus and grafana and want enriched full stack metrics

## When to avoid
- your workloads run on windows or macos since only linux is supported
- you need arm architecture support which is not available
- you want a hosted saas monitoring solution rather than a self-hosted agent
- your kubernetes cluster cannot run privileged daemonset agents required for ebpf

## Facets
- artifact type: application
- maturity: active
- function: monitoring, tracing, benchmarking
- domain: monitoring, cloud-computing
- platform: self-hosted, go
- tags: ebpf, kubernetes, trace-profiling, network-monitoring, prometheus, grafana, daemonset, apm, root-cause-analysis, devops, containers, linux, docker

## Member repositories
- harmonycloud/kindling (main) score 65

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:26.227233+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:56:44.825159+00:00, confidence not recorded.
  - readme: https://github.com/harmonycloud/kindling (fetched 2026-08-28T04:03:26.227233+00:00, sha 6843ecf12d37)
  - homepage: http://kindling.harmonycloud.cn (fetched 2026-08-29T12:58:25.652372+00:00, sha 9a1231c480dc)
  - site_page: http://kindling.harmonycloud.cn/docs/overview-and-concepts/overview (fetched 2026-08-29T12:58:25.661568+00:00, sha a542ddd9d98a)
  - site_page: http://kindling.harmonycloud.cn/docs/installation/kindling-agent/requirements (fetched 2026-08-29T12:58:25.663752+00:00, sha 1d103ea8037d)
  - site_page: http://kindling.harmonycloud.cn/docs/overview-and-concepts/trace-profiling (fetched 2026-08-29T12:58:25.665287+00:00, sha 36a5ddb211a3)
- Data as of 2026-08-30T08:39:29.467469+00:00.