Service level

Uptime + response targets.

We don't hide behind boilerplate. Below is the honest read on what to expect and what we're willing to put in writing.

Uptime targets

Component

Target

Observed

Marketing site (trytracelight.com)

99.9% monthly

Vercel edge — historically 99.99%+

Dashboard (authenticated)

99.9% monthly

Vercel + Supabase — historically 99.95%+

Public REST API (/api/v1)

99.9% monthly

Same infra as dashboard

Cron jobs (monitor runs, retention purge)

Best-effort within 15 min of scheduled time

Vercel Cron — typically <5min

Webhook dispatch (outbound)

Best-effort with 1 retry on 5xx

Median <2s; tail <30s

OSINT enrichment latency

<90s p95 for single subject

Median ~45s with all 32 workers running

Support response targets

Plan

First response

Resolution target

Starter

1 business day

Best-effort

Pro

Same business day

Best-effort

Agency

4 business hours

Same business day for sev-1

Enterprise

Per contract

SLA posture per plan

Starter / Pro:No formal contractual SLA. We ship the targets above as “our intent”; service credits aren't available at this tier. If the platform is down, we're working on it. The status page will reflect.
Agency: Soft SLA — same uptime targets with documented response posture. No service credits today; credits structure ships when we have first paying Agency customer.
Enterprise: Custom contractual SLA available. Service credits, mutual NDAs, custom DPAs all on the table. Email legal@trytracelight.com.

Honest disclosure on early-stage SLAs

Tracelight is a small operation. We can ship 99.9% uptime targets credibly because our infrastructure (Vercel + Supabase) historically does much better than that. We cannot ship a 99.99% SLA because at 4.3 minutes/month of downtime budget the math doesn't work for a small team.

If your procurement requires 99.99% with credits, the answer is currently “not yet — talk to us when we're bigger.” We won't lie about it.