observability
Create an alert on a metric
Changed onCreate an alert that fires when a metric crosses a threshold. Three sub_types:
- normal — Failure alert. Fires per failing call. Only for binary metrics (eval_type starts with binary_) plus the latency / infrastructure / critical-deviation built-ins.
- significant_change — Trend alert. Fires when a numeric/continuous metric's rolling mean deviates from history. Pass preset: "balanced" (or "conservative" / "aggressive") to skip manual std_multiplier / ewma_alpha / window_size math — the server expands the preset. Balanced is the safe default.
- threshold — Hard ceiling/floor. Fires when ≥ min_breaches of the last window_size evaluations cross value. Use for numeric metrics where the user has a known good range. Operator is gt | lt | outside ({low, high}).
Slack delivery requires notifications.slack.workspace_id (+ optional channel_id). Alerts with empty notifications are silent — fire counters still increment but no message is sent. Pick a workspace_id from slack_workspaces_list.
post/observability/v1/alerts/
Request
- Base URLhttps://api.cekura.ai
- URLhttps://api.cekura.ai/observability/v1/alerts/
- Auth
- API key in header X-CEKURA-API-KEY
- or HTTP bearer
- or HTTP bearer
- or HTTP bearer
Request body
Response
Changes
- info
endpoint added
- info