SLA Compliance Tracking

The Challenge

You've promised customers:

• 99.9% uptime
• Response times under 3 seconds
• Global availability

How do you prove it?

The Solution

Agent Status provides continuous monitoring data that feeds directly into SLA reporting.

Uptime Calculation

Uptime = (UP checks + REACHABLE checks + DEGRADED checks) / Total checks × 100

Only DOWN is counted as "down" for uptime purposes.

Example

168 checks in 7 days:
- UP: 150
- DEGRADED: 15
- DOWN: 3

Uptime = (150 + 15) / 168 = 98.2%

SLA Targets

SLA	Allowed Downtime/Month	Checks (hourly)	Max DOWN checks
99.9%	43 minutes	720	~1
99.5%	3.6 hours	720	3.6
99.0%	7.2 hours	720	7.2

Error Budget

Error budget = allowed downtime based on SLA.

SLA: 99.9%
Month: 30 days = 720 hours

Error budget: 720 × 0.1% = 0.72 hours = 43 minutes

Used this month: 15 minutes (1 incident)
Remaining: 28 minutes

When error budget runs out, freeze deployments and focus on reliability.

Customer-Facing Status Page

Public Status Badge

<script
  src="https://api.agentstatus.dev/static/widget/agentstatus.js"
  data-agent-id="your-agent-id"
></script>

Historical Uptime Display

Last 90 days:
████████████████████████████████████░██ 99.7%
                                    ↑
                              15 min outage
                              Feb 10, 2026

Latency SLA

If you've committed to response time SLAs:

SLA: P95 response < 3 seconds

This month:
- P95 achieved: 2.4s ✅
- SLA met: Yes

Configure TTFB SLA in Agent Status to track this automatically:

{
  "ttfb_sla_ms": 3000
}

Incident Post-Mortems

For each DOWN incident, Agent Status provides:

• When: Timestamp of status change
• Duration: Time until recovery
• Scope: Affected regions
• Error codes: What failed
• Recovery: When it came back