SLOs

The SLOController manages Service Level Objectives that define reliability targets for platform services. SLOs track error budgets, availability targets, and latency targets, providing a structured approach to reliability engineering.

ServiceLevelObjective Structure

Field	Type	Description
`id`	String	SLO identifier
`name`	String	SLO name (e.g., "AI Service Availability")
`description`	String	Description of the objective
`service`	String	Target service name
`sliType`	String	`availability`, `latency`, `error_rate`, `throughput`
`target`	double	Target value (e.g., 99.9 for 99.9% availability)
`window`	String	Evaluation window (`7d`, `28d`, `30d`)
`query`	String	PromQL query for the SLI measurement
`alertOnBudgetBurn`	boolean	Alert when error budget burn rate is high

SLO Management

Create SLO

Endpoint: POST /api/v1/observability/slos

curl -X POST http://localhost:8088/api/v1/observability/slos \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer ${TOKEN}" \
  -H "X-Tenant-ID: 550e8400" \
  -d '{
    "name": "AI Service Availability",
    "service": "ai-service",
    "sliType": "availability",
    "target": 99.9,
    "window": "30d",
    "alertOnBudgetBurn": true
  }'

List SLOs

Endpoint: GET /api/v1/observability/slos

Get SLO

Endpoint: GET /api/v1/observability/slos/:sloId

Update SLO

Endpoint: PUT /api/v1/observability/slos/:sloId

Delete SLO

Endpoint: DELETE /api/v1/observability/slos/:sloId

SLO Status

Endpoint: GET /api/v1/observability/slos/:sloId/status

Returns the current status of an SLO including:

SLOStatus Structure

Field	Type	Description
`sloId`	String	SLO identifier
`currentValue`	double	Current SLI value
`target`	double	SLO target
`errorBudgetTotal`	double	Total error budget for the window
`errorBudgetRemaining`	double	Remaining error budget
`errorBudgetConsumed`	double	Percentage of budget consumed
`burnRate`	double	Current error budget burn rate
`status`	String	`healthy`, `warning`, `breaching`
`timeWindowStart`	Instant	Start of evaluation window
`timeWindowEnd`	Instant	End of evaluation window

Error Budget

The error budget represents the acceptable amount of unreliability within the SLO window:

Error Budget = 1 - SLO Target
Example: 99.9% target = 0.1% error budget = ~43 minutes per 30 days

Budget Burn Rate Alerts

When alertOnBudgetBurn is enabled, alerts fire based on how fast the error budget is being consumed:

Burn Rate	Window	Severity	Description
14.4x	1 hour	Critical	Budget exhausted in 2 hours
6x	6 hours	Critical	Budget exhausted in 5 hours
3x	1 day	Warning	Budget exhausted in 10 days
1x	3 days	Info	Budget on track to be exhausted

SLO Dashboard

Endpoint: GET /api/v1/observability/slos/dashboard

Returns all SLOs with their current status for a consolidated SLO dashboard view.

Alerting Anomaly Detection