docs(alerting): initial info

docs/sources/alerting/alerting.md

@@ -6,9 +6,102 @@ page_keywords: alerting, grafana, plugins, documentation
 
 # Alerting
 
-`status of alerting`
+> Alerting is in very early development.
 
-Limitations.
-* no executor
-* only one alert per panel
+The roadmap for alerting is described in [issue #2209](https://github.com/grafana/grafana/issues/2209#issuecomment-210077445) and the current state can be found at this page.
 
+## Introduction
+
+So far Grafana does only support saving alering rules but not execute it. This means that any one using grafana for alerts have to extract the alerts from grafana and then import them into their own monitoring systems. The current defintion of an alert rule looks like this:
+
+``` go
+type AlertRule struct {
+  Id           int64  `json:"id"`
+  OrgId        int64  `json:"-"`
+  DashboardId  int64  `json:"dashboardId"`
+  PanelId      int64  `json:"panelId"`
+  Query        string `json:"query"`
+  QueryRefId   string `json:"queryRefId"`
+  WarnLevel    int64  `json:"warnLevel"`
+  CritLevel    int64  `json:"critLevel"`
+  WarnOperator string `json:"warnOperator"`
+  CritOperator string `json:"critOperator"`
+  Interval     string `json:"interval"`
+  Title        string `json:"title"`
+  Description  string `json:"description"`
+  QueryRange   string `json:"queryRange"`
+  Aggregator   string `json:"aggregator"`
+  State        string `json:"state"`
+}
+```
+
+Most of these properties might require some extra explaination.
+
+Query: json representation of the query used by grafana. Differes depending on datasource.
+QueryRange: The time range for which the query should look back.
+Aggregator: How the result should be reduced into a single value. ex avg, sum, min, max
+State: Current state of the alert OK, WARN, CRITICAL, ACKNOWLEGED.
+
+You can configure these settings in the Alerting tab on graph panels in edit mode. When the dashboard is saved the alert is created, updated based on what data you provide. If you wish to delete an alert you simply set the query to ' - select query - ' in the alerting tab and save the dashboard. (this will be improved within the UI later on).
+
+
+
+## Api
+
+Endpoints
+
+``` http
+GET /api/alerts/rules
+state //array of strings optional
+dashboardId //int optional
+panelId //int optional
+
+Result
+[]AlertRule
+```
+
+``` http
+GET /api/alerts/changes
+limit //array of strings optional
+sinceId //int optional
+
+Result
+[
+  {
+    id: incrementing id,
+    alertId: alertId,
+    type: CREATED/UPDATED/DELETED,
+    created: timestamp,
+  }
+]
+```
+
+``` http
+GET /api/alerts/rulres/:alertId
+
+Result AlertRule
+```
+
+``` http
+GET /api/alerts/rulres/:alertId/states
+
+Result
+[
+  {
+    alertId: int,
+    newState: OK, WARN, CRITICAL, ACKNOWLEGED,
+    created: timestamp,
+    info: description of what might have caused the changed alert state
+  }
+]
+```
+
+``` http
+PUT /api/alerts/rulres/:alertId/state
+Request
+{
+  alertId: alertid,
+  newState: OK, WARN, CRITICAL, ACKNOWLEGED,
+  info: description of what might have caused the changed alert state
+}
+```