mirror of
https://github.com/VictoriaMetrics/VictoriaMetrics.git
synced 2024-11-21 14:44:00 +00:00
Vmalert compliance improvements (#2320)
* vmalert: add support for `sortByLabel` template function * vmalert: update API according to Prometheus conformance program The changes to the API, field names and URL path has been made according to the Prometheus specification for `alert_generator` https://github.com/prometheus/compliance/blob/main/alert_generator/specification.md * vmalert: fix the timestamp of the evaluated rules The timestamp used for alert's `EndsAt` was calculated before sending the notification. While the correct way is to use the timestamp taken right before rules evaluation. * vmalert: add `-datasource.queryTimeAlignment` flag The flag is supposed to provide ability to disable `time` param alignment when executing rules. By default, this flag is enabled, so it remains backward compatible. The flag was introduced to achieve better compatibility with Prometheus behaviour according to https://github.com/prometheus/compliance/blob/main/alert_generator/specification.md Signed-off-by: hagen1778 <roman@victoriametrics.com>
This commit is contained in:
Roman Khavronenko
GitHub
parent
918ed5cb32
commit
fb6eab03a2
14 changed files with 590 additions and 582 deletions
|
|
@ -526,6 +526,8 @@ The shortlist of configuration flags is the following:
|
|||
Optional OAuth2 tokenURL to use for -datasource.url.
|
||||
-datasource.queryStep duration
|
||||
queryStep defines how far a value can fallback to when evaluating queries. For example, if datasource.queryStep=15s then param "step" with value "15s" will be added to every query.If queryStep isn't specified, rule's evaluationInterval will be used instead.
|
||||
-datasource.queryTimeAlignment
|
||||
Whether to align "time" parameter with evaluation interval.Alignment supposed to produce deterministic results despite of number of vmalert replicas or time they were started. See more details here https://github.com/VictoriaMetrics/VictoriaMetrics/pull/1257 (default true)
|
||||
-datasource.roundDigits int
|
||||
Adds "round_digits" GET param to datasource requests. In VM "round_digits" limits the number of digits after the decimal point in response values.
|
||||
-datasource.tlsCAFile string
|
||||
|
|
|
|||
|
|
@ -38,6 +38,8 @@ type AlertingRule struct {
|
|||
alerts map[uint64]*notifier.Alert
|
||||
// stores last moment of time Exec was called
|
||||
lastExecTime time.Time
|
||||
// stores the duration of the last Exec call
|
||||
lastExecDuration time.Duration
|
||||
// stores last error that happened in Exec func
|
||||
// resets on every successful Exec
|
||||
// may be used as Health state
|
||||
|
|
@ -203,12 +205,14 @@ func (ar *AlertingRule) ExecRange(ctx context.Context, start, end time.Time) ([]
|
|||
// Exec executes AlertingRule expression via the given Querier.
|
||||
// Based on the Querier results AlertingRule maintains notifier.Alerts
|
||||
func (ar *AlertingRule) Exec(ctx context.Context) ([]prompbmarshal.TimeSeries, error) {
|
||||
start := time.Now()
|
||||
qMetrics, err := ar.q.Query(ctx, ar.Expr)
|
||||
ar.mu.Lock()
|
||||
defer ar.mu.Unlock()
|
||||
|
||||
ar.lastExecTime = start
|
||||
ar.lastExecDuration = time.Since(start)
|
||||
ar.lastExecError = err
|
||||
ar.lastExecTime = time.Now()
|
||||
ar.lastExecSamples = len(qMetrics)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("failed to execute query %q: %w", ar.Expr, err)
|
||||
|
|
@ -386,31 +390,48 @@ func (ar *AlertingRule) AlertAPI(id uint64) *APIAlert {
|
|||
return ar.newAlertAPI(*a)
|
||||
}
|
||||
|
||||
// RuleAPI returns Rule representation in form
|
||||
// of APIAlertingRule
|
||||
func (ar *AlertingRule) RuleAPI() APIAlertingRule {
|
||||
var lastErr string
|
||||
// ToAPI returns Rule representation in form
|
||||
// of APIRule
|
||||
func (ar *AlertingRule) ToAPI() APIRule {
|
||||
r := APIRule{
|
||||
Type: "alerting",
|
||||
DatasourceType: ar.Type.String(),
|
||||
Name: ar.Name,
|
||||
Query: ar.Expr,
|
||||
Duration: ar.For.Seconds(),
|
||||
Labels: ar.Labels,
|
||||
Annotations: ar.Annotations,
|
||||
LastEvaluation: ar.lastExecTime,
|
||||
EvaluationTime: ar.lastExecDuration.Seconds(),
|
||||
Health: "ok",
|
||||
State: "inactive",
|
||||
Alerts: ar.AlertsToAPI(),
|
||||
LastSamples: ar.lastExecSamples,
|
||||
|
||||
// encode as strings to avoid rounding in JSON
|
||||
ID: fmt.Sprintf("%d", ar.ID()),
|
||||
GroupID: fmt.Sprintf("%d", ar.GroupID),
|
||||
}
|
||||
if ar.lastExecError != nil {
|
||||
lastErr = ar.lastExecError.Error()
|
||||
r.LastError = ar.lastExecError.Error()
|
||||
r.Health = "err"
|
||||
}
|
||||
return APIAlertingRule{
|
||||
// encode as strings to avoid rounding
|
||||
ID: fmt.Sprintf("%d", ar.ID()),
|
||||
GroupID: fmt.Sprintf("%d", ar.GroupID),
|
||||
Type: ar.Type.String(),
|
||||
Name: ar.Name,
|
||||
Expression: ar.Expr,
|
||||
For: ar.For.String(),
|
||||
LastError: lastErr,
|
||||
LastSamples: ar.lastExecSamples,
|
||||
LastExec: ar.lastExecTime,
|
||||
Labels: ar.Labels,
|
||||
Annotations: ar.Annotations,
|
||||
// satisfy APIRule.State logic
|
||||
if len(r.Alerts) > 0 {
|
||||
r.State = notifier.StatePending.String()
|
||||
stateFiring := notifier.StateFiring.String()
|
||||
for _, a := range r.Alerts {
|
||||
if a.State == stateFiring {
|
||||
r.State = stateFiring
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
return r
|
||||
}
|
||||
|
||||
// AlertsAPI generates list of APIAlert objects from existing alerts
|
||||
func (ar *AlertingRule) AlertsAPI() []*APIAlert {
|
||||
// AlertsToAPI generates list of APIAlert objects from existing alerts
|
||||
func (ar *AlertingRule) AlertsToAPI() []*APIAlert {
|
||||
var alerts []*APIAlert
|
||||
ar.mu.RLock()
|
||||
for _, a := range ar.alerts {
|
||||
|
|
|
|||
|
|
@ -38,6 +38,8 @@ var (
|
|||
queryStep = flag.Duration("datasource.queryStep", 0, "queryStep defines how far a value can fallback to when evaluating queries. "+
|
||||
"For example, if datasource.queryStep=15s then param \"step\" with value \"15s\" will be added to every query."+
|
||||
"If queryStep isn't specified, rule's evaluationInterval will be used instead.")
|
||||
queryTimeAlignment = flag.Bool("datasource.queryTimeAlignment", true, `Whether to align "time" parameter with evaluation interval.`+
|
||||
"Alignment supposed to produce deterministic results despite of number of vmalert replicas or time they were started. See more details here https://github.com/VictoriaMetrics/VictoriaMetrics/pull/1257")
|
||||
maxIdleConnections = flag.Int("datasource.maxIdleConnections", 100, `Defines the number of idle (keep-alive connections) to each configured datasource. Consider setting this value equal to the value: groups_total * group.concurrency. Too low a value may result in a high number of sockets in TIME_WAIT state.`)
|
||||
roundDigits = flag.Int("datasource.roundDigits", 0, `Adds "round_digits" GET param to datasource requests. `+
|
||||
`In VM "round_digits" limits the number of digits after the decimal point in response values.`)
|
||||
|
|
|
|||
|
|
@ -125,7 +125,7 @@ func (s *VMStorage) setPrometheusInstantReqParams(r *http.Request, query string,
|
|||
if s.lookBack > 0 {
|
||||
timestamp = timestamp.Add(-s.lookBack)
|
||||
}
|
||||
if s.evaluationInterval > 0 {
|
||||
if *queryTimeAlignment && s.evaluationInterval > 0 {
|
||||
// see https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1232
|
||||
timestamp = timestamp.Truncate(s.evaluationInterval)
|
||||
}
|
||||
|
|
|
|||
|
|
@ -19,14 +19,15 @@ import (
|
|||
|
||||
// Group is an entity for grouping rules
|
||||
type Group struct {
|
||||
mu sync.RWMutex
|
||||
Name string
|
||||
File string
|
||||
Rules []Rule
|
||||
Type datasource.Type
|
||||
Interval time.Duration
|
||||
Concurrency int
|
||||
Checksum string
|
||||
mu sync.RWMutex
|
||||
Name string
|
||||
File string
|
||||
Rules []Rule
|
||||
Type datasource.Type
|
||||
Interval time.Duration
|
||||
Concurrency int
|
||||
Checksum string
|
||||
LastEvaluation time.Time
|
||||
|
||||
Labels map[string]string
|
||||
Params url.Values
|
||||
|
|
@ -283,6 +284,7 @@ func (g *Group) start(ctx context.Context, nts func() []notifier.Notifier, rw *r
|
|||
logger.Errorf("group %q: %s", g.Name, err)
|
||||
}
|
||||
}
|
||||
g.LastEvaluation = iterationStart
|
||||
}
|
||||
g.metrics.iterationDuration.UpdateDuration(iterationStart)
|
||||
}
|
||||
|
|
@ -347,6 +349,7 @@ var (
|
|||
func (e *executor) exec(ctx context.Context, rule Rule, resolveDuration time.Duration) error {
|
||||
execTotal.Inc()
|
||||
|
||||
now := time.Now()
|
||||
tss, err := rule.Exec(ctx)
|
||||
if err != nil {
|
||||
execErrors.Inc()
|
||||
|
|
@ -371,12 +374,12 @@ func (e *executor) exec(ctx context.Context, rule Rule, resolveDuration time.Dur
|
|||
for _, a := range ar.alerts {
|
||||
switch a.State {
|
||||
case notifier.StateFiring:
|
||||
a.End = time.Now().Add(resolveDuration)
|
||||
a.End = now.Add(resolveDuration)
|
||||
alerts = append(alerts, *a)
|
||||
case notifier.StateInactive:
|
||||
// set End to execStart to notify
|
||||
// that it was just resolved
|
||||
a.End = time.Now()
|
||||
a.End = now
|
||||
alerts = append(alerts, *a)
|
||||
}
|
||||
}
|
||||
|
|
|
|||
|
|
@ -163,21 +163,17 @@ func (g *Group) toAPI() APIGroup {
|
|||
// encode as string to avoid rounding
|
||||
ID: fmt.Sprintf("%d", g.ID()),
|
||||
|
||||
Name: g.Name,
|
||||
Type: g.Type.String(),
|
||||
File: g.File,
|
||||
Interval: g.Interval.String(),
|
||||
Concurrency: g.Concurrency,
|
||||
Params: urlValuesToStrings(g.Params),
|
||||
Labels: g.Labels,
|
||||
Name: g.Name,
|
||||
Type: g.Type.String(),
|
||||
File: g.File,
|
||||
Interval: g.Interval.Seconds(),
|
||||
LastEvaluation: g.LastEvaluation,
|
||||
Concurrency: g.Concurrency,
|
||||
Params: urlValuesToStrings(g.Params),
|
||||
Labels: g.Labels,
|
||||
}
|
||||
for _, r := range g.Rules {
|
||||
switch v := r.(type) {
|
||||
case *AlertingRule:
|
||||
ag.AlertingRules = append(ag.AlertingRules, v.RuleAPI())
|
||||
case *RecordingRule:
|
||||
ag.RecordingRules = append(ag.RecordingRules, v.RuleAPI())
|
||||
}
|
||||
ag.Rules = append(ag.Rules, r.ToAPI())
|
||||
}
|
||||
return ag
|
||||
}
|
||||
|
|
|
|||
|
|
@ -20,6 +20,7 @@ import (
|
|||
"net"
|
||||
"net/url"
|
||||
"regexp"
|
||||
"sort"
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
|
|
@ -282,6 +283,14 @@ func InitTemplateFunc(externalURL *url.URL) {
|
|||
return m.Labels[label]
|
||||
},
|
||||
|
||||
// sortByLabel sorts the given metrics by provided label key
|
||||
"sortByLabel": func(label string, metrics []metric) []metric {
|
||||
sort.SliceStable(metrics, func(i, j int) bool {
|
||||
return metrics[i].Labels[label] < metrics[j].Labels[label]
|
||||
})
|
||||
return metrics
|
||||
},
|
||||
|
||||
// value returns the value of the given metric.
|
||||
// usually used alongside with `query` template function.
|
||||
"value": func(m metric) float64 {
|
||||
|
|
|
|||
|
|
@ -31,6 +31,8 @@ type RecordingRule struct {
|
|||
mu sync.RWMutex
|
||||
// stores last moment of time Exec was called
|
||||
lastExecTime time.Time
|
||||
// stores the duration of the last Exec call
|
||||
lastExecDuration time.Duration
|
||||
// stores last error that happened in Exec func
|
||||
// resets on every successful Exec
|
||||
// may be used as Health state
|
||||
|
|
@ -123,11 +125,13 @@ func (rr *RecordingRule) ExecRange(ctx context.Context, start, end time.Time) ([
|
|||
|
||||
// Exec executes RecordingRule expression via the given Querier.
|
||||
func (rr *RecordingRule) Exec(ctx context.Context) ([]prompbmarshal.TimeSeries, error) {
|
||||
start := time.Now()
|
||||
qMetrics, err := rr.q.Query(ctx, rr.Expr)
|
||||
rr.mu.Lock()
|
||||
defer rr.mu.Unlock()
|
||||
|
||||
rr.lastExecTime = time.Now()
|
||||
rr.lastExecTime = start
|
||||
rr.lastExecDuration = time.Since(start)
|
||||
rr.lastExecError = err
|
||||
rr.lastExecSamples = len(qMetrics)
|
||||
if err != nil {
|
||||
|
|
@ -193,23 +197,27 @@ func (rr *RecordingRule) UpdateWith(r Rule) error {
|
|||
return nil
|
||||
}
|
||||
|
||||
// RuleAPI returns Rule representation in form
|
||||
// of APIRecordingRule
|
||||
func (rr *RecordingRule) RuleAPI() APIRecordingRule {
|
||||
var lastErr string
|
||||
if rr.lastExecError != nil {
|
||||
lastErr = rr.lastExecError.Error()
|
||||
}
|
||||
return APIRecordingRule{
|
||||
// ToAPI returns Rule's representation in form
|
||||
// of APIRule
|
||||
func (rr *RecordingRule) ToAPI() APIRule {
|
||||
r := APIRule{
|
||||
Type: "recording",
|
||||
DatasourceType: rr.Type.String(),
|
||||
Name: rr.Name,
|
||||
Query: rr.Expr,
|
||||
Labels: rr.Labels,
|
||||
LastEvaluation: rr.lastExecTime,
|
||||
EvaluationTime: rr.lastExecDuration.Seconds(),
|
||||
Health: "ok",
|
||||
LastSamples: rr.lastExecSamples,
|
||||
// encode as strings to avoid rounding
|
||||
ID: fmt.Sprintf("%d", rr.ID()),
|
||||
GroupID: fmt.Sprintf("%d", rr.GroupID),
|
||||
Name: rr.Name,
|
||||
Type: rr.Type.String(),
|
||||
Expression: rr.Expr,
|
||||
LastError: lastErr,
|
||||
LastSamples: rr.lastExecSamples,
|
||||
LastExec: rr.lastExecTime,
|
||||
Labels: rr.Labels,
|
||||
ID: fmt.Sprintf("%d", rr.ID()),
|
||||
GroupID: fmt.Sprintf("%d", rr.GroupID),
|
||||
}
|
||||
|
||||
if rr.lastExecError != nil {
|
||||
r.LastError = rr.lastExecError.Error()
|
||||
r.Health = "err"
|
||||
}
|
||||
return r
|
||||
}
|
||||
|
|
|
|||
|
|
@ -21,6 +21,8 @@ type Rule interface {
|
|||
// UpdateWith performs modification of current Rule
|
||||
// with fields of the given Rule.
|
||||
UpdateWith(Rule) error
|
||||
// ToAPI converts Rule into APIRule
|
||||
ToAPI() APIRule
|
||||
// Close performs the shutdown procedures for rule
|
||||
// such as metrics unregister
|
||||
Close()
|
||||
|
|
|
|||
|
|
@ -29,7 +29,7 @@ func initLinks() {
|
|||
pathPrefix = "/"
|
||||
}
|
||||
apiLinks = [][2]string{
|
||||
{path.Join(pathPrefix, "api/v1/groups"), "list all loaded groups and rules"},
|
||||
{path.Join(pathPrefix, "api/v1/rules"), "list all loaded groups and rules"},
|
||||
{path.Join(pathPrefix, "api/v1/alerts"), "list all active alerts"},
|
||||
{path.Join(pathPrefix, "api/v1/groupID/alertID/status"), "get alert status by ID"},
|
||||
{path.Join(pathPrefix, "flags"), "command-line flags"},
|
||||
|
|
@ -75,7 +75,7 @@ func (rh *requestHandler) handler(w http.ResponseWriter, r *http.Request) bool {
|
|||
case "/notifiers":
|
||||
WriteListTargets(w, notifier.GetTargets())
|
||||
return true
|
||||
case "/api/v1/groups":
|
||||
case "/api/v1/rules":
|
||||
data, err := rh.listGroups()
|
||||
if err != nil {
|
||||
httpserver.Errorf(w, r, "%s", err)
|
||||
|
|
@ -127,10 +127,10 @@ func (rh *requestHandler) handler(w http.ResponseWriter, r *http.Request) bool {
|
|||
}
|
||||
|
||||
type listGroupsResponse struct {
|
||||
Data struct {
|
||||
Status string `json:"status"`
|
||||
Data struct {
|
||||
Groups []APIGroup `json:"groups"`
|
||||
} `json:"data"`
|
||||
Status string `json:"status"`
|
||||
}
|
||||
|
||||
func (rh *requestHandler) groups() []APIGroup {
|
||||
|
|
@ -149,6 +149,7 @@ func (rh *requestHandler) groups() []APIGroup {
|
|||
|
||||
return groups
|
||||
}
|
||||
|
||||
func (rh *requestHandler) listGroups() ([]byte, error) {
|
||||
lr := listGroupsResponse{Status: "success"}
|
||||
lr.Data.Groups = rh.groups()
|
||||
|
|
@ -163,10 +164,10 @@ func (rh *requestHandler) listGroups() ([]byte, error) {
|
|||
}
|
||||
|
||||
type listAlertsResponse struct {
|
||||
Data struct {
|
||||
Status string `json:"status"`
|
||||
Data struct {
|
||||
Alerts []*APIAlert `json:"alerts"`
|
||||
} `json:"data"`
|
||||
Status string `json:"status"`
|
||||
}
|
||||
|
||||
func (rh *requestHandler) groupAlerts() []GroupAlerts {
|
||||
|
|
@ -181,7 +182,7 @@ func (rh *requestHandler) groupAlerts() []GroupAlerts {
|
|||
if !ok {
|
||||
continue
|
||||
}
|
||||
alerts = append(alerts, a.AlertsAPI()...)
|
||||
alerts = append(alerts, a.AlertsToAPI()...)
|
||||
}
|
||||
if len(alerts) > 0 {
|
||||
groupAlerts = append(groupAlerts, GroupAlerts{
|
||||
|
|
@ -204,7 +205,7 @@ func (rh *requestHandler) listAlerts() ([]byte, error) {
|
|||
if !ok {
|
||||
continue
|
||||
}
|
||||
lr.Data.Alerts = append(lr.Data.Alerts, a.AlertsAPI()...)
|
||||
lr.Data.Alerts = append(lr.Data.Alerts, a.AlertsToAPI()...)
|
||||
}
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -31,14 +31,7 @@
|
|||
rOk := make(map[string]int)
|
||||
rNotOk := make(map[string]int)
|
||||
for _, g := range groups {
|
||||
for _, r := range g.AlertingRules{
|
||||
if r.LastError != "" {
|
||||
rNotOk[g.Name]++
|
||||
} else {
|
||||
rOk[g.Name]++
|
||||
}
|
||||
}
|
||||
for _, r := range g.RecordingRules{
|
||||
for _, r := range g.Rules {
|
||||
if r.LastError != "" {
|
||||
rNotOk[g.Name]++
|
||||
} else {
|
||||
|
|
@ -52,7 +45,7 @@
|
|||
{% for _, g := range groups %}
|
||||
<div class="group-heading{% if rNotOk[g.Name] > 0 %} alert-danger{% endif %}" data-bs-target="rules-{%s g.ID %}">
|
||||
<span class="anchor" id="group-{%s g.ID %}"></span>
|
||||
<a href="#group-{%s g.ID %}">{%s g.Name %}{% if g.Type != "prometheus" %} ({%s g.Type %}){% endif %} (every {%s g.Interval %})</a>
|
||||
<a href="#group-{%s g.ID %}">{%s g.Name %}{% if g.Type != "prometheus" %} ({%s g.Type %}){% endif %} (every {%f.0 g.Interval %}s)</a>
|
||||
{% if rNotOk[g.Name] > 0 %}<span class="badge bg-danger" title="Number of rules with status Error">{%d rNotOk[g.Name] %}</span> {% endif %}
|
||||
<span class="badge bg-success" title="Number of rules withs status Ok">{%d rOk[g.Name] %}</span>
|
||||
<p class="fs-6 fw-lighter">{%s g.File %}</p>
|
||||
|
|
@ -75,34 +68,24 @@
|
|||
</tr>
|
||||
</thead>
|
||||
<tbody>
|
||||
{% for _, ar := range g.AlertingRules %}
|
||||
<tr{% if ar.LastError != "" %} class="alert-danger"{% endif %}>
|
||||
{% for _, r := range g.Rules %}
|
||||
<tr{% if r.LastError != "" %} class="alert-danger"{% endif %}>
|
||||
<td>
|
||||
<b>alert:</b> {%s ar.Name %} (for: {%v ar.For %})<br>
|
||||
<code><pre>{%s ar.Expression %}</pre></code><br>
|
||||
{% if len(ar.Labels) > 0 %} <b>Labels:</b>{% endif %}
|
||||
{% for k, v := range ar.Labels %}
|
||||
{% if r.Type == "alerting" %}
|
||||
<b>alert:</b> (for: {%v r.Duration %})
|
||||
{% else %}
|
||||
<b>record:</b> {%s r.Name %}
|
||||
{% endif %}
|
||||
<br>
|
||||
<code><pre>{%s r.Query %}</pre></code><br>
|
||||
{% if len(r.Labels) > 0 %} <b>Labels:</b>{% endif %}
|
||||
{% for k, v := range r.Labels %}
|
||||
<span class="ms-1 badge bg-primary">{%s k %}={%s v %}</span>
|
||||
{% endfor %}
|
||||
</td>
|
||||
<td><div class="error-cell">{%s ar.LastError %}</div></td>
|
||||
<td>{%d ar.LastSamples %}</td>
|
||||
<td>{%f.3 time.Since(ar.LastExec).Seconds() %}s ago</td>
|
||||
</tr>
|
||||
{% endfor %}
|
||||
{% for _, rr := range g.RecordingRules %}
|
||||
<tr>
|
||||
<td>
|
||||
<b>record:</b> {%s rr.Name %}<br>
|
||||
<code><pre>{%s rr.Expression %}</pre></code>
|
||||
{% if len(rr.Labels) > 0 %} <b>Labels:</b>{% endif %}
|
||||
{% for k, v := range rr.Labels %}
|
||||
<span class="ms-1 badge bg-primary">{%s k %}={%s v %}</span>
|
||||
{% endfor %}
|
||||
</td>
|
||||
<td><div class="error-cell">{%s rr.LastError %}</div></td>
|
||||
<td>{%d rr.LastSamples %}</td>
|
||||
<td>{%f.3 time.Since(rr.LastExec).Seconds() %}s ago</td>
|
||||
<td><div class="error-cell">{%s r.LastError %}</div></td>
|
||||
<td>{%d r.LastSamples %}</td>
|
||||
<td>{%f.3 time.Since(r.LastEvaluation).Seconds() %}s ago</td>
|
||||
</tr>
|
||||
{% endfor %}
|
||||
</tbody>
|
||||
|
|
|
|||
File diff suppressed because it is too large
Load diff
|
|
@ -54,9 +54,9 @@ func TestHandler(t *testing.T) {
|
|||
t.Errorf("expected 1 alert got %d", length)
|
||||
}
|
||||
})
|
||||
t.Run("/api/v1/groups", func(t *testing.T) {
|
||||
t.Run("/api/v1/rules", func(t *testing.T) {
|
||||
lr := listGroupsResponse{}
|
||||
getResp(ts.URL+"/api/v1/groups", &lr, 200)
|
||||
getResp(ts.URL+"/api/v1/rules", &lr, 200)
|
||||
if length := len(lr.Data.Groups); length != 1 {
|
||||
t.Errorf("expected 1 group got %d", length)
|
||||
}
|
||||
|
|
|
|||
|
|
@ -4,63 +4,61 @@ import (
|
|||
"time"
|
||||
)
|
||||
|
||||
// APIAlert represents an notifier.AlertingRule state
|
||||
// APIAlert represents a notifier.AlertingRule state
|
||||
// for WEB view
|
||||
// https://github.com/prometheus/compliance/blob/main/alert_generator/specification.md#get-apiv1rules
|
||||
type APIAlert struct {
|
||||
ID string `json:"id"`
|
||||
Name string `json:"name"`
|
||||
RuleID string `json:"rule_id"`
|
||||
GroupID string `json:"group_id"`
|
||||
Expression string `json:"expression"`
|
||||
State string `json:"state"`
|
||||
Name string `json:"name"`
|
||||
Value string `json:"value"`
|
||||
Labels map[string]string `json:"labels"`
|
||||
Labels map[string]string `json:"labels,omitempty"`
|
||||
Annotations map[string]string `json:"annotations"`
|
||||
ActiveAt time.Time `json:"activeAt"`
|
||||
SourceLink string `json:"source"`
|
||||
Restored bool `json:"restored"`
|
||||
|
||||
// Additional fields
|
||||
|
||||
// ID is an unique Alert's ID within a group
|
||||
ID string `json:"id"`
|
||||
// RuleID is an unique Rule's ID within a group
|
||||
RuleID string `json:"rule_id"`
|
||||
// GroupID is an unique Group's ID
|
||||
GroupID string `json:"group_id"`
|
||||
// Expression contains the PromQL/MetricsQL expression
|
||||
// for Rule's evaluation
|
||||
Expression string `json:"expression"`
|
||||
// SourceLink contains a link to a system which should show
|
||||
// why Alert was generated
|
||||
SourceLink string `json:"source"`
|
||||
// Restored shows whether Alert's state was restored on restart
|
||||
Restored bool `json:"restored"`
|
||||
}
|
||||
|
||||
// APIGroup represents Group for WEB view
|
||||
// https://github.com/prometheus/compliance/blob/main/alert_generator/specification.md#get-apiv1rules
|
||||
type APIGroup struct {
|
||||
Name string `json:"name"`
|
||||
Type string `json:"type"`
|
||||
ID string `json:"id"`
|
||||
File string `json:"file"`
|
||||
Interval string `json:"interval"`
|
||||
Concurrency int `json:"concurrency"`
|
||||
Params []string `json:"params"`
|
||||
Labels map[string]string `json:"labels,omitempty"`
|
||||
AlertingRules []APIAlertingRule `json:"alerting_rules"`
|
||||
RecordingRules []APIRecordingRule `json:"recording_rules"`
|
||||
}
|
||||
// Name is the group name as present in the config
|
||||
Name string `json:"name"`
|
||||
// Rules contains both recording and alerting rules
|
||||
Rules []APIRule `json:"rules"`
|
||||
// Interval is the Group's evaluation interval in float seconds as present in the file.
|
||||
Interval float64 `json:"interval"`
|
||||
// LastEvaluation is the timestamp of the last time the Group was executed
|
||||
LastEvaluation time.Time `json:"lastEvaluation"`
|
||||
|
||||
// APIAlertingRule represents AlertingRule for WEB view
|
||||
type APIAlertingRule struct {
|
||||
ID string `json:"id"`
|
||||
Name string `json:"name"`
|
||||
Type string `json:"type"`
|
||||
GroupID string `json:"group_id"`
|
||||
Expression string `json:"expression"`
|
||||
For string `json:"for"`
|
||||
LastError string `json:"last_error"`
|
||||
LastSamples int `json:"last_samples"`
|
||||
LastExec time.Time `json:"last_exec"`
|
||||
Labels map[string]string `json:"labels"`
|
||||
Annotations map[string]string `json:"annotations"`
|
||||
}
|
||||
// Additional fields
|
||||
|
||||
// APIRecordingRule represents RecordingRule for WEB view
|
||||
type APIRecordingRule struct {
|
||||
ID string `json:"id"`
|
||||
Name string `json:"name"`
|
||||
Type string `json:"type"`
|
||||
GroupID string `json:"group_id"`
|
||||
Expression string `json:"expression"`
|
||||
LastError string `json:"last_error"`
|
||||
LastSamples int `json:"last_samples"`
|
||||
LastExec time.Time `json:"last_exec"`
|
||||
Labels map[string]string `json:"labels"`
|
||||
// Type shows the datasource type (prometheus or graphite) of the Group
|
||||
Type string `json:"type"`
|
||||
// ID is a unique Group ID
|
||||
ID string `json:"id"`
|
||||
// File contains a path to the file with Group's config
|
||||
File string `json:"file"`
|
||||
// Concurrency shows how many rules may be evaluated simultaneously
|
||||
Concurrency int `json:"concurrency"`
|
||||
// Params contains HTTP URL parameters added to each Rule's request
|
||||
Params []string `json:"params,omitempty"`
|
||||
// Labels is a set of label value pairs, that will be added to every rule.
|
||||
Labels map[string]string `json:"labels,omitempty"`
|
||||
}
|
||||
|
||||
// GroupAlerts represents a group of alerts for WEB view
|
||||
|
|
@ -68,3 +66,43 @@ type GroupAlerts struct {
|
|||
Group APIGroup
|
||||
Alerts []*APIAlert
|
||||
}
|
||||
|
||||
// APIRule represents a Rule for WEB view
|
||||
// see https://github.com/prometheus/compliance/blob/main/alert_generator/specification.md#get-apiv1rules
|
||||
type APIRule struct {
|
||||
// State must be one of these under following scenarios
|
||||
// "pending": at least 1 alert in the rule in pending state and no other alert in firing state.
|
||||
// "firing": at least 1 alert in the rule in firing state.
|
||||
// "inactive": no alert in the rule in firing or pending state.
|
||||
State string `json:"state"`
|
||||
Name string `json:"name"`
|
||||
// Query represents Rule's `expression` field
|
||||
Query string `json:"query"`
|
||||
// Duration represents Rule's `for` field
|
||||
Duration float64 `json:"duration"`
|
||||
Labels map[string]string `json:"labels,omitempty"`
|
||||
Annotations map[string]string `json:"annotations,omitempty"`
|
||||
// LastError contains the error faced while executing the rule.
|
||||
LastError string `json:"lastError"`
|
||||
// EvaluationTime is the time taken to completely evaluate the rule in float seconds.
|
||||
EvaluationTime float64 `json:"evaluationTime"`
|
||||
// LastEvaluation is the timestamp of the last time the rule was executed
|
||||
LastEvaluation time.Time `json:"lastEvaluation"`
|
||||
// Alerts is the list of all the alerts in this rule that are currently pending or firing
|
||||
Alerts []*APIAlert `json:"alerts,omitempty"`
|
||||
// Health is the health of rule evaluation.
|
||||
// It MUST be one of "ok", "err", "unknown"
|
||||
Health string `json:"health"`
|
||||
// Type of the rule: recording or alerting
|
||||
Type string `json:"type"`
|
||||
|
||||
// Additional fields
|
||||
|
||||
// Type of the rule: recording or alerting
|
||||
DatasourceType string `json:"datasourceType"`
|
||||
LastSamples int `json:"lastSamples"`
|
||||
// ID is an unique Alert's ID within a group
|
||||
ID string `json:"id"`
|
||||
// GroupID is an unique Group's ID
|
||||
GroupID string `json:"group_id"`
|
||||
}
|
||||
|
|
|
|||
Loading…
Reference in a new issue