mirror of
https://github.com/grafana/grafana.git
synced 2025-02-25 18:55:37 -06:00
Alerting: Enhancements to /rules (#33085)
* set processing time * merge labels and set on response * use state cache for adding alerts to rules * minor cleanup * pr feedback * Do not initialize mutex unnecessarily Co-authored-by: Ganesh Vernekar <15064823+codesome@users.noreply.github.com> * linter Co-authored-by: Ganesh Vernekar <15064823+codesome@users.noreply.github.com>
This commit is contained in:
parent
7480d9e2be
commit
4be1d84f23
@ -5,6 +5,7 @@ import (
|
||||
"net/http"
|
||||
"time"
|
||||
|
||||
"github.com/grafana/grafana/pkg/services/ngalert/eval"
|
||||
apiv1 "github.com/prometheus/client_golang/api/prometheus/v1"
|
||||
|
||||
ngmodels "github.com/grafana/grafana/pkg/services/ngalert/models"
|
||||
@ -84,18 +85,9 @@ func (srv PrometheusSrv) RouteGetRuleStatuses(c *models.ReqContext) response.Res
|
||||
LastEvaluation: time.Time{},
|
||||
EvaluationTime: 0, // TODO: see if we are able to pass this along with evaluation results
|
||||
}
|
||||
for _, rule := range alertRuleQuery.Result {
|
||||
instanceQuery := ngmodels.ListAlertInstancesQuery{
|
||||
DefinitionOrgID: c.SignedInUser.OrgId,
|
||||
DefinitionUID: rule.UID,
|
||||
}
|
||||
if err := srv.store.ListAlertInstances(&instanceQuery); err != nil {
|
||||
ruleResponse.DiscoveryBase.Status = "error"
|
||||
ruleResponse.DiscoveryBase.Error = fmt.Sprintf("failure getting alerts for rule %s: %s", rule.UID, err.Error())
|
||||
ruleResponse.DiscoveryBase.ErrorType = apiv1.ErrServer
|
||||
return response.JSON(http.StatusInternalServerError, ruleResponse)
|
||||
}
|
||||
|
||||
stateMap := srv.stateTracker.GetStatesByRuleUID()
|
||||
for _, rule := range alertRuleQuery.Result {
|
||||
alertingRule := apimodels.AlertingRule{
|
||||
State: "inactive",
|
||||
Name: rule.Title,
|
||||
@ -106,36 +98,46 @@ func (srv PrometheusSrv) RouteGetRuleStatuses(c *models.ReqContext) response.Res
|
||||
|
||||
newRule := apimodels.Rule{
|
||||
Name: rule.Title,
|
||||
Labels: nil, // TODO: NG AlertRule does not have labels but does have annotations
|
||||
Labels: rule.Labels,
|
||||
Health: "ok", // TODO: update this in the future when error and noData states are being evaluated and set
|
||||
Type: apiv1.RuleTypeAlerting,
|
||||
LastEvaluation: time.Time{}, // TODO: set this to be rule evaluation time once it is being set
|
||||
EvaluationTime: 0, // TODO: set this once we are saving it or adding it to evaluation results
|
||||
LastEvaluation: time.Time{},
|
||||
}
|
||||
for _, instance := range instanceQuery.Result {
|
||||
activeAt := instance.CurrentStateSince
|
||||
|
||||
for _, alertState := range stateMap[rule.UID] {
|
||||
activeAt := alertState.StartsAt
|
||||
alert := &apimodels.Alert{
|
||||
Labels: map[string]string(instance.Labels),
|
||||
Annotations: nil, // TODO: set these once they are added to evaluation results
|
||||
State: translateInstanceState(instance.CurrentState),
|
||||
Labels: map[string]string(alertState.Labels),
|
||||
Annotations: alertState.Annotations,
|
||||
State: alertState.State.String(),
|
||||
ActiveAt: &activeAt,
|
||||
Value: "", // TODO: set this once it is added to the evaluation results
|
||||
}
|
||||
if instance.LastEvalTime.After(newRule.LastEvaluation) {
|
||||
newRule.LastEvaluation = instance.LastEvalTime
|
||||
newGroup.LastEvaluation = instance.LastEvalTime
|
||||
|
||||
if alertState.LastEvaluationTime.After(newRule.LastEvaluation) {
|
||||
newRule.LastEvaluation = alertState.LastEvaluationTime
|
||||
newGroup.LastEvaluation = alertState.LastEvaluationTime
|
||||
}
|
||||
switch alert.State {
|
||||
case "pending":
|
||||
|
||||
alertingRule.Duration = alertState.EvaluationDuration.Seconds()
|
||||
newRule.EvaluationTime = alertState.EvaluationDuration.Seconds()
|
||||
|
||||
switch alertState.State {
|
||||
case eval.Normal:
|
||||
case eval.Pending:
|
||||
if alertingRule.State == "inactive" {
|
||||
alertingRule.State = "pending"
|
||||
}
|
||||
case "firing":
|
||||
case eval.Alerting:
|
||||
alertingRule.State = "firing"
|
||||
case eval.Error:
|
||||
// handle Error case based on configuration in alertRule
|
||||
case eval.NoData:
|
||||
// handle NoData case based on configuration in alertRule
|
||||
}
|
||||
|
||||
alertingRule.Alerts = append(alertingRule.Alerts, alert)
|
||||
}
|
||||
|
||||
alertingRule.Rule = newRule
|
||||
newGroup.Rules = append(newGroup.Rules, alertingRule)
|
||||
newGroup.Interval = float64(rule.IntervalSeconds)
|
||||
@ -144,16 +146,3 @@ func (srv PrometheusSrv) RouteGetRuleStatuses(c *models.ReqContext) response.Res
|
||||
}
|
||||
return response.JSON(http.StatusOK, ruleResponse)
|
||||
}
|
||||
|
||||
func translateInstanceState(state ngmodels.InstanceStateType) string {
|
||||
switch {
|
||||
case state == ngmodels.InstanceStateFiring:
|
||||
return "firing"
|
||||
case state == ngmodels.InstanceStateNormal:
|
||||
return "inactive"
|
||||
case state == ngmodels.InstanceStatePending:
|
||||
return "pending"
|
||||
default:
|
||||
return "inactive"
|
||||
}
|
||||
}
|
||||
|
@ -76,6 +76,11 @@ const (
|
||||
// that evaluated to true (Alerting).
|
||||
Alerting
|
||||
|
||||
// Pending is the eval state for an alert instance condition
|
||||
// that evaluated to true (Alerting) but has not yet met
|
||||
// the For duration defined in AlertRule
|
||||
Pending
|
||||
|
||||
// NoData is the eval state for an alert rule condition
|
||||
// that evaluated to NoData.
|
||||
NoData
|
||||
@ -86,7 +91,7 @@ const (
|
||||
)
|
||||
|
||||
func (s State) String() string {
|
||||
return [...]string{"Normal", "Alerting", "NoData", "Error"}[s]
|
||||
return [...]string{"Normal", "Alerting", "Pending", "NoData", "Error"}[s]
|
||||
}
|
||||
|
||||
// AlertExecCtx is the context provided for executing an alert condition.
|
||||
|
@ -78,7 +78,7 @@ func (sch *schedule) ruleRoutine(grafanaCtx context.Context, key models.AlertRul
|
||||
return err
|
||||
}
|
||||
|
||||
processedStates := stateTracker.ProcessEvalResults(alertRule, results)
|
||||
processedStates := stateTracker.ProcessEvalResults(alertRule, results, end.Sub(start))
|
||||
sch.saveAlertStates(processedStates)
|
||||
alerts := FromAlertStateToPostableAlerts(processedStates)
|
||||
sch.log.Debug("sending alerts to notifier", "count", len(alerts.PostableAlerts))
|
||||
@ -321,7 +321,7 @@ func (sch *schedule) saveAlertStates(states []state.AlertState) {
|
||||
for _, s := range states {
|
||||
cmd := models.SaveAlertInstanceCommand{
|
||||
DefinitionOrgID: s.OrgID,
|
||||
DefinitionUID: s.UID,
|
||||
DefinitionUID: s.AlertRuleUID,
|
||||
Labels: models.InstanceLabels(s.Labels),
|
||||
State: models.InstanceStateType(s.State.String()),
|
||||
LastEvalTime: s.LastEvaluationTime,
|
||||
@ -330,7 +330,7 @@ func (sch *schedule) saveAlertStates(states []state.AlertState) {
|
||||
}
|
||||
err := sch.store.SaveAlertInstance(&cmd)
|
||||
if err != nil {
|
||||
sch.log.Error("failed to save alert state", "uid", s.UID, "orgId", s.OrgID, "labels", s.Labels.String(), "state", s.State.String(), "msg", err.Error())
|
||||
sch.log.Error("failed to save alert state", "uid", s.AlertRuleUID, "orgId", s.OrgID, "labels", s.Labels.String(), "state", s.State.String(), "msg", err.Error())
|
||||
}
|
||||
}
|
||||
}
|
||||
@ -355,7 +355,7 @@ func (sch *schedule) WarmStateCache(st *state.StateTracker) {
|
||||
for _, entry := range cmd.Result {
|
||||
lbs := map[string]string(entry.Labels)
|
||||
stateForEntry := state.AlertState{
|
||||
UID: entry.DefinitionUID,
|
||||
AlertRuleUID: entry.DefinitionUID,
|
||||
OrgID: entry.DefinitionOrgID,
|
||||
CacheId: fmt.Sprintf("%s %s", entry.DefinitionUID, lbs),
|
||||
Labels: lbs,
|
||||
|
@ -5,6 +5,8 @@ import (
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
prometheusModel "github.com/prometheus/common/model"
|
||||
|
||||
"github.com/grafana/grafana/pkg/infra/log"
|
||||
|
||||
"github.com/grafana/grafana-plugin-sdk-go/data"
|
||||
@ -13,7 +15,7 @@ import (
|
||||
)
|
||||
|
||||
type AlertState struct {
|
||||
UID string
|
||||
AlertRuleUID string
|
||||
OrgID int64
|
||||
CacheId string
|
||||
Labels data.Labels
|
||||
@ -22,6 +24,7 @@ type AlertState struct {
|
||||
StartsAt time.Time
|
||||
EndsAt time.Time
|
||||
LastEvaluationTime time.Time
|
||||
EvaluationDuration time.Duration
|
||||
Annotations map[string]string
|
||||
}
|
||||
|
||||
@ -31,21 +34,20 @@ type StateEvaluation struct {
|
||||
}
|
||||
|
||||
type cache struct {
|
||||
cacheMap map[string]AlertState
|
||||
mu sync.Mutex
|
||||
states map[string]AlertState
|
||||
mtxStates sync.Mutex
|
||||
}
|
||||
|
||||
type StateTracker struct {
|
||||
stateCache cache
|
||||
quit chan struct{}
|
||||
Log log.Logger
|
||||
cache cache
|
||||
quit chan struct{}
|
||||
Log log.Logger
|
||||
}
|
||||
|
||||
func NewStateTracker(logger log.Logger) *StateTracker {
|
||||
tracker := &StateTracker{
|
||||
stateCache: cache{
|
||||
cacheMap: make(map[string]AlertState),
|
||||
mu: sync.Mutex{},
|
||||
cache: cache{
|
||||
states: make(map[string]AlertState),
|
||||
},
|
||||
quit: make(chan struct{}),
|
||||
Log: logger,
|
||||
@ -54,135 +56,158 @@ func NewStateTracker(logger log.Logger) *StateTracker {
|
||||
return tracker
|
||||
}
|
||||
|
||||
func (st *StateTracker) getOrCreate(alertRule *ngModels.AlertRule, result eval.Result) AlertState {
|
||||
st.stateCache.mu.Lock()
|
||||
defer st.stateCache.mu.Unlock()
|
||||
lbs := data.Labels{}
|
||||
if len(result.Instance) > 0 {
|
||||
lbs = result.Instance
|
||||
}
|
||||
func (st *StateTracker) getOrCreate(alertRule *ngModels.AlertRule, result eval.Result, evaluationDuration time.Duration) AlertState {
|
||||
st.cache.mtxStates.Lock()
|
||||
defer st.cache.mtxStates.Unlock()
|
||||
|
||||
// if duplicate labels exist, alertRule label will take precedence
|
||||
lbs := mergeLabels(alertRule.Labels, result.Instance)
|
||||
lbs["__alert_rule_uid__"] = alertRule.UID
|
||||
lbs["__alert_rule_namespace_uid__"] = alertRule.NamespaceUID
|
||||
lbs["__alert_rule_title__"] = alertRule.Title
|
||||
lbs[prometheusModel.AlertNameLabel] = alertRule.Title
|
||||
|
||||
id := fmt.Sprintf("%s", map[string]string(lbs))
|
||||
if state, ok := st.cache.states[id]; ok {
|
||||
return state
|
||||
}
|
||||
|
||||
annotations := map[string]string{}
|
||||
if len(alertRule.Annotations) > 0 {
|
||||
annotations = alertRule.Annotations
|
||||
}
|
||||
|
||||
idString := fmt.Sprintf("%s", map[string]string(lbs))
|
||||
if state, ok := st.stateCache.cacheMap[idString]; ok {
|
||||
return state
|
||||
newResults := []StateEvaluation{
|
||||
{
|
||||
EvaluationTime: result.EvaluatedAt,
|
||||
EvaluationState: result.State,
|
||||
},
|
||||
}
|
||||
|
||||
st.Log.Debug("adding new alert state cache entry", "cacheId", idString, "state", result.State.String(), "evaluatedAt", result.EvaluatedAt.String())
|
||||
st.Log.Debug("adding new alert state cache entry", "cacheId", id, "state", result.State.String(), "evaluatedAt", result.EvaluatedAt.String())
|
||||
newState := AlertState{
|
||||
UID: alertRule.UID,
|
||||
OrgID: alertRule.OrgID,
|
||||
CacheId: idString,
|
||||
Labels: lbs,
|
||||
State: result.State,
|
||||
Results: []StateEvaluation{},
|
||||
Annotations: annotations,
|
||||
AlertRuleUID: alertRule.UID,
|
||||
OrgID: alertRule.OrgID,
|
||||
CacheId: id,
|
||||
Labels: lbs,
|
||||
State: result.State,
|
||||
Results: newResults,
|
||||
Annotations: annotations,
|
||||
EvaluationDuration: evaluationDuration,
|
||||
}
|
||||
if result.State == eval.Alerting {
|
||||
newState.StartsAt = result.EvaluatedAt
|
||||
}
|
||||
st.stateCache.cacheMap[idString] = newState
|
||||
st.cache.states[id] = newState
|
||||
return newState
|
||||
}
|
||||
|
||||
func (st *StateTracker) set(stateEntry AlertState) {
|
||||
st.stateCache.mu.Lock()
|
||||
defer st.stateCache.mu.Unlock()
|
||||
st.stateCache.cacheMap[stateEntry.CacheId] = stateEntry
|
||||
func (st *StateTracker) set(entry AlertState) {
|
||||
st.cache.mtxStates.Lock()
|
||||
defer st.cache.mtxStates.Unlock()
|
||||
st.cache.states[entry.CacheId] = entry
|
||||
}
|
||||
|
||||
func (st *StateTracker) Get(stateId string) AlertState {
|
||||
st.stateCache.mu.Lock()
|
||||
defer st.stateCache.mu.Unlock()
|
||||
return st.stateCache.cacheMap[stateId]
|
||||
func (st *StateTracker) Get(id string) AlertState {
|
||||
st.cache.mtxStates.Lock()
|
||||
defer st.cache.mtxStates.Unlock()
|
||||
return st.cache.states[id]
|
||||
}
|
||||
|
||||
//Used to ensure a clean cache on startup
|
||||
func (st *StateTracker) ResetCache() {
|
||||
st.stateCache.mu.Lock()
|
||||
defer st.stateCache.mu.Unlock()
|
||||
st.stateCache.cacheMap = make(map[string]AlertState)
|
||||
st.cache.mtxStates.Lock()
|
||||
defer st.cache.mtxStates.Unlock()
|
||||
st.cache.states = make(map[string]AlertState)
|
||||
}
|
||||
|
||||
func (st *StateTracker) ProcessEvalResults(alertRule *ngModels.AlertRule, results eval.Results) []AlertState {
|
||||
func (st *StateTracker) ProcessEvalResults(alertRule *ngModels.AlertRule, results eval.Results, evaluationDuration time.Duration) []AlertState {
|
||||
st.Log.Info("state tracker processing evaluation results", "uid", alertRule.UID, "resultCount", len(results))
|
||||
var changedStates []AlertState
|
||||
var states []AlertState
|
||||
for _, result := range results {
|
||||
s, _ := st.setNextState(alertRule, result)
|
||||
changedStates = append(changedStates, s)
|
||||
s := st.setNextState(alertRule, result, evaluationDuration)
|
||||
states = append(states, s)
|
||||
}
|
||||
st.Log.Debug("returning changed states to scheduler", "count", len(changedStates))
|
||||
return changedStates
|
||||
st.Log.Debug("returning changed states to scheduler", "count", len(states))
|
||||
return states
|
||||
}
|
||||
|
||||
//TODO: When calculating if an alert should not be firing anymore, we should take three things into account:
|
||||
// 1. The re-send the delay if any, we don't want to send every firing alert every time, we should have a fixed delay across all alerts to avoid saturating the notification system
|
||||
// 2. The evaluation interval defined for this particular alert - we don't support that yet but will eventually allow you to define how often do you want this alert to be evaluted
|
||||
// 3. The base interval defined by the scheduler - in the case where #2 is not yet an option we can use the base interval at which every alert runs.
|
||||
//Set the current state based on evaluation results
|
||||
//return the state and a bool indicating whether a state transition occurred
|
||||
func (st *StateTracker) setNextState(alertRule *ngModels.AlertRule, result eval.Result) (AlertState, bool) {
|
||||
currentState := st.getOrCreate(alertRule, result)
|
||||
func (st *StateTracker) setNextState(alertRule *ngModels.AlertRule, result eval.Result, evaluationDuration time.Duration) AlertState {
|
||||
currentState := st.getOrCreate(alertRule, result, evaluationDuration)
|
||||
st.Log.Debug("setting alert state", "uid", alertRule.UID)
|
||||
switch {
|
||||
case currentState.State == result.State:
|
||||
st.Log.Debug("no state transition", "cacheId", currentState.CacheId, "state", currentState.State.String())
|
||||
currentState.LastEvaluationTime = result.EvaluatedAt
|
||||
currentState.EvaluationDuration = evaluationDuration
|
||||
currentState.Results = append(currentState.Results, StateEvaluation{
|
||||
EvaluationTime: result.EvaluatedAt,
|
||||
EvaluationState: result.State,
|
||||
})
|
||||
if currentState.State == eval.Alerting {
|
||||
currentState.EndsAt = result.EvaluatedAt.Add(40 * time.Second)
|
||||
currentState.EndsAt = result.EvaluatedAt.Add(alertRule.For * time.Second)
|
||||
}
|
||||
st.set(currentState)
|
||||
return currentState, false
|
||||
return currentState
|
||||
case currentState.State == eval.Normal && result.State == eval.Alerting:
|
||||
st.Log.Debug("state transition from normal to alerting", "cacheId", currentState.CacheId)
|
||||
currentState.State = eval.Alerting
|
||||
currentState.LastEvaluationTime = result.EvaluatedAt
|
||||
currentState.StartsAt = result.EvaluatedAt
|
||||
currentState.EndsAt = result.EvaluatedAt.Add(40 * time.Second)
|
||||
currentState.EndsAt = result.EvaluatedAt.Add(alertRule.For * time.Second)
|
||||
currentState.EvaluationDuration = evaluationDuration
|
||||
currentState.Results = append(currentState.Results, StateEvaluation{
|
||||
EvaluationTime: result.EvaluatedAt,
|
||||
EvaluationState: result.State,
|
||||
})
|
||||
currentState.Annotations["alerting"] = result.EvaluatedAt.String()
|
||||
st.set(currentState)
|
||||
return currentState, true
|
||||
return currentState
|
||||
case currentState.State == eval.Alerting && result.State == eval.Normal:
|
||||
st.Log.Debug("state transition from alerting to normal", "cacheId", currentState.CacheId)
|
||||
currentState.State = eval.Normal
|
||||
currentState.LastEvaluationTime = result.EvaluatedAt
|
||||
currentState.EndsAt = result.EvaluatedAt
|
||||
currentState.EvaluationDuration = evaluationDuration
|
||||
currentState.Results = append(currentState.Results, StateEvaluation{
|
||||
EvaluationTime: result.EvaluatedAt,
|
||||
EvaluationState: result.State,
|
||||
})
|
||||
st.set(currentState)
|
||||
return currentState, true
|
||||
return currentState
|
||||
default:
|
||||
return currentState, false
|
||||
return currentState
|
||||
}
|
||||
}
|
||||
|
||||
func (st *StateTracker) GetAll() []AlertState {
|
||||
var states []AlertState
|
||||
st.stateCache.mu.Lock()
|
||||
defer st.stateCache.mu.Unlock()
|
||||
for _, v := range st.stateCache.cacheMap {
|
||||
st.cache.mtxStates.Lock()
|
||||
defer st.cache.mtxStates.Unlock()
|
||||
for _, v := range st.cache.states {
|
||||
states = append(states, v)
|
||||
}
|
||||
return states
|
||||
}
|
||||
|
||||
func (st *StateTracker) GetStatesByRuleUID() map[string][]AlertState {
|
||||
ruleMap := make(map[string][]AlertState)
|
||||
st.cache.mtxStates.Lock()
|
||||
defer st.cache.mtxStates.Unlock()
|
||||
for _, state := range st.cache.states {
|
||||
if ruleStates, ok := ruleMap[state.AlertRuleUID]; ok {
|
||||
ruleStates = append(ruleStates, state)
|
||||
ruleMap[state.AlertRuleUID] = ruleStates
|
||||
} else {
|
||||
ruleStates := []AlertState{state}
|
||||
ruleMap[state.AlertRuleUID] = ruleStates
|
||||
}
|
||||
}
|
||||
return ruleMap
|
||||
}
|
||||
|
||||
func (st *StateTracker) cleanUp() {
|
||||
ticker := time.NewTicker(time.Duration(60) * time.Minute)
|
||||
st.Log.Debug("starting cleanup process", "intervalMinutes", 60)
|
||||
@ -200,9 +225,9 @@ func (st *StateTracker) cleanUp() {
|
||||
|
||||
func (st *StateTracker) trim() {
|
||||
st.Log.Info("trimming alert state cache", "now", time.Now())
|
||||
st.stateCache.mu.Lock()
|
||||
defer st.stateCache.mu.Unlock()
|
||||
for _, v := range st.stateCache.cacheMap {
|
||||
st.cache.mtxStates.Lock()
|
||||
defer st.cache.mtxStates.Unlock()
|
||||
for _, v := range st.cache.states {
|
||||
if len(v.Results) > 100 {
|
||||
st.Log.Debug("trimming result set", "cacheId", v.CacheId, "count", len(v.Results)-100)
|
||||
newResults := make([]StateEvaluation, 100)
|
||||
@ -214,7 +239,7 @@ func (st *StateTracker) trim() {
|
||||
}
|
||||
|
||||
func (a AlertState) Equals(b AlertState) bool {
|
||||
return a.UID == b.UID &&
|
||||
return a.AlertRuleUID == b.AlertRuleUID &&
|
||||
a.OrgID == b.OrgID &&
|
||||
a.CacheId == b.CacheId &&
|
||||
a.Labels.String() == b.Labels.String() &&
|
||||
@ -229,3 +254,17 @@ func (st *StateTracker) Put(states []AlertState) {
|
||||
st.set(s)
|
||||
}
|
||||
}
|
||||
|
||||
// if duplicate labels exist, keep the value from the first set
|
||||
func mergeLabels(a, b data.Labels) data.Labels {
|
||||
newLbs := data.Labels{}
|
||||
for k, v := range a {
|
||||
newLbs[k] = v
|
||||
}
|
||||
for k, v := range b {
|
||||
if _, ok := newLbs[k]; !ok {
|
||||
newLbs[k] = v
|
||||
}
|
||||
}
|
||||
return newLbs
|
||||
}
|
||||
|
@ -35,11 +35,11 @@ func TestWarmStateCache(t *testing.T) {
|
||||
|
||||
expectedEntries := []state.AlertState{
|
||||
{
|
||||
UID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: "test_uid map[test1:testValue1]",
|
||||
Labels: data.Labels{"test1": "testValue1"},
|
||||
State: eval.Normal,
|
||||
AlertRuleUID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: "test_uid map[test1:testValue1]",
|
||||
Labels: data.Labels{"test1": "testValue1"},
|
||||
State: eval.Normal,
|
||||
Results: []state.StateEvaluation{
|
||||
{EvaluationTime: evaluationTime, EvaluationState: eval.Normal},
|
||||
},
|
||||
@ -47,11 +47,11 @@ func TestWarmStateCache(t *testing.T) {
|
||||
EndsAt: evaluationTime.Add(1 * time.Minute),
|
||||
LastEvaluationTime: evaluationTime,
|
||||
}, {
|
||||
UID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: "test_uid map[test2:testValue2]",
|
||||
Labels: data.Labels{"test2": "testValue2"},
|
||||
State: eval.Alerting,
|
||||
AlertRuleUID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: "test_uid map[test2:testValue2]",
|
||||
Labels: data.Labels{"test2": "testValue2"},
|
||||
State: eval.Alerting,
|
||||
Results: []state.StateEvaluation{
|
||||
{EvaluationTime: evaluationTime, EvaluationState: eval.Alerting},
|
||||
},
|
||||
|
@ -20,7 +20,11 @@ func TestProcessEvalResults(t *testing.T) {
|
||||
if err != nil {
|
||||
t.Fatalf("error parsing date format: %s", err.Error())
|
||||
}
|
||||
cacheId := "map[__alert_rule_namespace_uid__:test_namespace __alert_rule_title__:test_title __alert_rule_uid__:test_uid label1:value1 label2:value2]"
|
||||
cacheId := "map[__alert_rule_namespace_uid__:test_namespace __alert_rule_uid__:test_uid alertname:test_title label1:value1 label2:value2 rule_label:rule_value]"
|
||||
|
||||
ruleLabels := map[string]string{
|
||||
"rule_label": "rule_value",
|
||||
}
|
||||
alertRule := models.AlertRule{
|
||||
ID: 1,
|
||||
OrgID: 123,
|
||||
@ -28,13 +32,17 @@ func TestProcessEvalResults(t *testing.T) {
|
||||
Condition: "A",
|
||||
UID: "test_uid",
|
||||
NamespaceUID: "test_namespace",
|
||||
For: 10 * time.Second,
|
||||
Labels: ruleLabels,
|
||||
}
|
||||
processingTime := 10 * time.Millisecond
|
||||
expectedLabels := data.Labels{
|
||||
"label1": "value1",
|
||||
"label2": "value2",
|
||||
"rule_label": "rule_value",
|
||||
"__alert_rule_uid__": "test_uid",
|
||||
"__alert_rule_namespace_uid__": "test_namespace",
|
||||
"__alert_rule_title__": "test_title",
|
||||
"alertname": "test_title",
|
||||
}
|
||||
testCases := []struct {
|
||||
desc string
|
||||
@ -60,11 +68,11 @@ func TestProcessEvalResults(t *testing.T) {
|
||||
expectedResultCount: 1,
|
||||
expectedCacheEntries: []state.AlertState{
|
||||
{
|
||||
UID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: cacheId,
|
||||
Labels: expectedLabels,
|
||||
State: eval.Normal,
|
||||
AlertRuleUID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: cacheId,
|
||||
Labels: expectedLabels,
|
||||
State: eval.Normal,
|
||||
Results: []state.StateEvaluation{
|
||||
{EvaluationTime: evaluationTime, EvaluationState: eval.Normal},
|
||||
},
|
||||
@ -94,17 +102,17 @@ func TestProcessEvalResults(t *testing.T) {
|
||||
expectedResultCount: 2,
|
||||
expectedCacheEntries: []state.AlertState{
|
||||
{
|
||||
UID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: cacheId,
|
||||
Labels: expectedLabels,
|
||||
State: eval.Alerting,
|
||||
AlertRuleUID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: cacheId,
|
||||
Labels: expectedLabels,
|
||||
State: eval.Alerting,
|
||||
Results: []state.StateEvaluation{
|
||||
{EvaluationTime: evaluationTime, EvaluationState: eval.Normal},
|
||||
{EvaluationTime: evaluationTime.Add(1 * time.Minute), EvaluationState: eval.Alerting},
|
||||
},
|
||||
StartsAt: evaluationTime.Add(1 * time.Minute),
|
||||
EndsAt: evaluationTime.Add(100 * time.Second),
|
||||
EndsAt: evaluationTime.Add(alertRule.For * time.Second).Add(1 * time.Minute),
|
||||
LastEvaluationTime: evaluationTime.Add(1 * time.Minute),
|
||||
},
|
||||
},
|
||||
@ -129,11 +137,11 @@ func TestProcessEvalResults(t *testing.T) {
|
||||
expectedResultCount: 2,
|
||||
expectedCacheEntries: []state.AlertState{
|
||||
{
|
||||
UID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: cacheId,
|
||||
Labels: expectedLabels,
|
||||
State: eval.Normal,
|
||||
AlertRuleUID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: cacheId,
|
||||
Labels: expectedLabels,
|
||||
State: eval.Normal,
|
||||
Results: []state.StateEvaluation{
|
||||
{EvaluationTime: evaluationTime, EvaluationState: eval.Alerting},
|
||||
{EvaluationTime: evaluationTime.Add(1 * time.Minute), EvaluationState: eval.Normal},
|
||||
@ -164,17 +172,17 @@ func TestProcessEvalResults(t *testing.T) {
|
||||
expectedResultCount: 2,
|
||||
expectedCacheEntries: []state.AlertState{
|
||||
{
|
||||
UID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: cacheId,
|
||||
Labels: expectedLabels,
|
||||
State: eval.Alerting,
|
||||
AlertRuleUID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: cacheId,
|
||||
Labels: expectedLabels,
|
||||
State: eval.Alerting,
|
||||
Results: []state.StateEvaluation{
|
||||
{EvaluationTime: evaluationTime, EvaluationState: eval.Alerting},
|
||||
{EvaluationTime: evaluationTime.Add(1 * time.Minute), EvaluationState: eval.Alerting},
|
||||
},
|
||||
StartsAt: evaluationTime,
|
||||
EndsAt: evaluationTime.Add(100 * time.Second),
|
||||
EndsAt: evaluationTime.Add(alertRule.For * time.Second).Add(1 * time.Minute),
|
||||
LastEvaluationTime: evaluationTime.Add(1 * time.Minute),
|
||||
},
|
||||
},
|
||||
@ -199,11 +207,11 @@ func TestProcessEvalResults(t *testing.T) {
|
||||
expectedResultCount: 2,
|
||||
expectedCacheEntries: []state.AlertState{
|
||||
{
|
||||
UID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: cacheId,
|
||||
Labels: expectedLabels,
|
||||
State: eval.Normal,
|
||||
AlertRuleUID: "test_uid",
|
||||
OrgID: 123,
|
||||
CacheId: cacheId,
|
||||
Labels: expectedLabels,
|
||||
State: eval.Normal,
|
||||
Results: []state.StateEvaluation{
|
||||
{EvaluationTime: evaluationTime, EvaluationState: eval.Normal},
|
||||
{EvaluationTime: evaluationTime.Add(1 * time.Minute), EvaluationState: eval.Normal},
|
||||
@ -219,7 +227,7 @@ func TestProcessEvalResults(t *testing.T) {
|
||||
for _, tc := range testCases {
|
||||
t.Run("all fields for a cache entry are set correctly", func(t *testing.T) {
|
||||
st := state.NewStateTracker(log.New("test_state_tracker"))
|
||||
_ = st.ProcessEvalResults(&alertRule, tc.evalResults)
|
||||
_ = st.ProcessEvalResults(&alertRule, tc.evalResults, processingTime)
|
||||
for _, entry := range tc.expectedCacheEntries {
|
||||
if !entry.Equals(st.Get(entry.CacheId)) {
|
||||
t.Log(tc.desc)
|
||||
@ -231,7 +239,7 @@ func TestProcessEvalResults(t *testing.T) {
|
||||
|
||||
t.Run("the expected number of entries are added to the cache", func(t *testing.T) {
|
||||
st := state.NewStateTracker(log.New("test_state_tracker"))
|
||||
st.ProcessEvalResults(&alertRule, tc.evalResults)
|
||||
st.ProcessEvalResults(&alertRule, tc.evalResults, processingTime)
|
||||
assert.Equal(t, len(tc.expectedCacheEntries), len(st.GetAll()))
|
||||
})
|
||||
|
||||
@ -240,15 +248,15 @@ func TestProcessEvalResults(t *testing.T) {
|
||||
//for a unique set of labels.
|
||||
t.Run("the expected number of states are returned to the caller", func(t *testing.T) {
|
||||
st := state.NewStateTracker(log.New("test_state_tracker"))
|
||||
results := st.ProcessEvalResults(&alertRule, tc.evalResults)
|
||||
results := st.ProcessEvalResults(&alertRule, tc.evalResults, processingTime)
|
||||
assert.Equal(t, len(tc.evalResults), len(results))
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
func printEntryDiff(a, b state.AlertState, t *testing.T) {
|
||||
if a.UID != b.UID {
|
||||
t.Log(fmt.Sprintf("%v \t %v\n", a.UID, b.UID))
|
||||
if a.AlertRuleUID != b.AlertRuleUID {
|
||||
t.Log(fmt.Sprintf("%v \t %v\n", a.AlertRuleUID, b.AlertRuleUID))
|
||||
}
|
||||
if a.OrgID != b.OrgID {
|
||||
t.Log(fmt.Sprintf("%v \t %v\n", a.OrgID, b.OrgID))
|
||||
|
Loading…
Reference in New Issue
Block a user