用户影响加权的SLO仪表盘。 通过实际客户暴露来衡量真实的99.99%正常运行时间——而不仅仅是原始可用性。

---

原始可用性的问题

每个SRE都知道公式：(total_time − downtime) / total_time。 但这个指标将影响3%用户的凌晨3点故障视为 中午停电影响所有高峰负载用户。

当值班平台显示99.99%时，是否意味着你的客户体验 99.99%的可用性？还是仅仅是你的基础设施99.99%的时间都在运行？

慢冲击回答了真正的问题。

---

建筑

原始的正常运行时间每分钟都很重要。影响的正常运行时间权重是谁受影响，以及是否是高峰流量。

演示仪表盘

从回购内演示收集器（合成银行平台，没有实时客户数据）中渲染出来。 以docker-compose up或slo-impact开启，打开http://localhost:8080.

它是如何运作的

对于每个事件，我们捕捉两个额外的维度：

每次事件用户影响停机时间：

slo-impact

User-impact-weighted SLO dashboard.
Measure true 99.99% uptime by actual customer exposure — not just raw availability.

---

The problem with raw availability

Every SRE knows the formula: (total_time − downtime) / total_time.
But that metric treats a 3 AM outage affecting 3% of users identically to a noon outage affecting all users at peak load.

When your oncall platform showed 99.99%, did it mean your customers experienced 99.99% availability? Or just that your infrastructure was up 99.99% of the time?

slo-impact answers the real question.

---

Architecture

Raw uptime counts every minute. Impact uptime weights who was affected and whether it was peak traffic.

Demo dashboard

Rendered from the in-repo demo collector (synthetic banking platform, no live customer data). Start it with docker-compose up or slo-impact and open http://localhost:8080.

How it works

For each incident we capture two extra dimensions:

User-impact downtime per incident:

impact_seconds = duration × user_fraction × traffic_weight

User-impact uptime:

user_impact_uptime = 1 − (Σ impact_seconds) / window_seconds

A 3 AM outage at 8% traffic affecting 5% of users contributes only 0.4% of what the same outage would contribute at peak. That gap — between what your infrastructure reports and what customers feel — is the number that matters for SLO integrity.

“If your SLO has never been breached, it isn’t an SLO. It’s a slogan.”

---

Quick start

Option 1 — Docker (recommended, zero setup)

git clone https://github.com/Ajay150313/slo-impact.git
cd slo-impact
docker-compose up

Open http://localhost:8080 — the dashboard runs with built-in synthetic data that mirrors a realistic digital banking platform. No Prometheus required.

Option 2 — pip

pip install slo-impact          # once published to PyPI
# or from source:
pip install -e .

slo-impact                      # starts on http://0.0.0.0:8080

Option 3 — from source (development)

git clone https://github.com/Ajay150313/slo-impact.git
cd slo-impact
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
uvicorn slo_impact.main:app --reload --port 8080

---

Prometheus integration