Create an SLO in Datadog: POST /api/v1/slo with type metric or monitor, defining numerator and denominator metric queries for metric-based SLOs or monitor IDs for monitor-based SLOs
Create a burn rate alert monitor: POST /api/v1/monitor with type slo alert and a query of the form error_budget(SLO_ID).over(1h) * 100 > BURN_RATE_THRESHOLD
Configure multi-window alerting by creating two monitors: a fast-burn monitor (e.g., 1h window, 14.4x burn rate) and a slow-burn monitor (e.g., 24h window, 1x burn rate) and combine them with a composite monitor
Set notification routing in the monitor message field using @pagerduty-SERVICE_NAME or @slack-CHANNEL to route to on-call
Enable monitor evaluation renotification to prevent silence periods when an SLO remains in burn
Retrieve SLO history: GET /api/v1/slo/{slo_id}/history?from_ts=START&to_ts=END to pull error budget consumption history for reporting
Known gotchas
Datadog SLO burn rate monitors are a paid feature; verify your Datadog plan includes SLO monitoring before configuring — enabling it on a plan that does not include it does not produce an error but monitors will not evaluate
Metric-based SLOs use 30-day rolling windows by default; the window affects burn rate calculation — ensure your SLO definition's target window matches the alerting timeframe
Composite monitors combine burn rate monitors with AND logic; if one constituent monitor is in no-data state, the composite may not alert as expected — configure no-data alerts on constituent monitors separately
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp