delimit_obs_alerts
Configure automated alerts for production thresholds such as latency and error rate by listing, creating, updating, or deleting alert rules based on metric series.
Instructions
Manage alerting rules — list, create, update, delete (experimental).
When to use: to configure ongoing alerts for production thresholds (latency, error rate, saturation, queue depth) against the same metric series visible via delimit_obs_metrics. Sub-actions: "list" inventories existing rules, "create" mints one, "update" edits, "delete" removes. When NOT to use: for one-shot metric queries (delimit_obs_metrics), log search (delimit_obs_logs), or the health rollup (delimit_obs_status). Also: do not call "create" repeatedly to retry a failed alert delivery — alerting is configuration, not delivery.
Sibling contrast: delimit_obs_metrics queries data; this configures automated thresholds against that data. Compared to cloud-provider alerting consoles, this routes through the ops bridge so the rule set is recorded in the same observability layer as the metric source.
Side effects: WRITES to the alert configuration on the ops
backend for action in ("create", "update", "delete"); reads only
for "list". Routes through backends.ops_bridge.obs_alerts. Marked
EXPERIMENTAL — the schema for alert_rule is backend-specific
and may evolve; pin tested rule shapes if depending on this in
production. No license gate at this level (gating handled by
the backend's own admin checks).
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| action | Yes | Alert sub-action. One of "list", "create", "update", "delete". Required. | |
| alert_rule | No | Rule definition dict (required for create/update). Backend-specific schema. | |
| rule_id | No | Identifier for an existing rule (required for delete/update). |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||