Command Center

指挥中心

Live operating picture for Northwind Labs - 12 products, one control plane.

Morning brief

AI generatedclaude-4.5-sonnet / 2 min read

Portfolio MRR is $432.8K, tracking +6.6% month over month. AI spend now absorbs 11.6% of revenue - the fastest-growing line in the cost base. There are 3 open incidents, and Voice Studio has the weakest health score in the portfolio - worth a look before standup.

Monthly recurring revenue

$432.8K

+6.6%vs last month
ARR $5.2M
Gross profit

$359.5K

+9.1%margin vs target
83.1% gross margin
AI spend this month

$50.3K

+18.4%vs last month
11.6% of revenue
Active users

198.9K

+9.2%vs last month
19.3K DAU / 116.2K MAU
Platform uptime

99.81%

+0.1%30-day average
Blended churn

3.51%

-0.4%revenue weighted
Deploys (30d)

310

+12.6%vs previous 30d
Open incidents

3

+40%vs yesterday
Avg health score 84/100
Portfolio trend
Aggregate across all products, last 90 days
Revenue mix
MRR contribution by product
Daily active users
Portfolio-wide, last 30 days

Needs attention

Lowest health scores in the portfolio

Live activity

Latest signals across every platform

View all

Code Review AI AI spend up 34% in 3 days

HighNew

Daily token cost moved from $93 to $125. Most of the increase comes from claude-4.5-opus retries on long-context requests.

15h ago

New error group in Resume AI: TypeError on /api/v1/export

MediumNew

First seen 11 minutes ago, 285 users affected across 6 releases. Culprit traced to Edge Functions.

16h ago

Recorder AI v4.8.2 deployed to production

Low

Build finished in 123s from northwind-labs/recorder@main. 16 files changed, no failing checks.

1d ago

API key rate anomaly detected

Critical

Key sk_live_…7584 issued 7,043 requests/min, 18x its 30-day baseline. The key has been rate-limited pending review.

1d ago

Fallback route activated for Translate Pro

Medium

gemini-3-flash returned 66 consecutive 529s. Traffic shifted to qwen-3-max for 17 minutes with no user-visible errors.

1d ago

Marketing OS p95 latency above the 3s threshold

High

p95 held at 2462ms for 12 minutes on ai-inference. Autoscaling added 2 replicas and latency is recovering.

2d ago