mirror of https://github.com/VictoriaMetrics/VictoriaMetrics.git synced 2024-12-15 08:23:34 +01:00

Aliaksandr Valialkin db6dadf1f7

docs: convert png images to webp in all the docs except of docs/operator/*

This reduces the size of docs/* folder from 33MB to 18MB

Images inside docs/operator/* must be converted at the https://github.com/VictoriaMetrics/operator/tree/master/docs
and then the updated images must be automatically propagated to the docs/operator/*

This is a follow-up for d3f919df3e

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/pull/5206

2023-11-22 19:29:47 +02:00

9.7 KiB

Raw Blame History

sort

weight

title

aliases

Alerting with vmalert and Managed VictoriaMetrics

docs

parent	weight
managed	4

/managed-victoriametrics/alerting-vmalert-managed-victoria-metrics.html

Alerting with vmalert and Managed VictoriaMetrics

This guide explains the different ways in which you can use vmalert in conjunction with Managed VictoriaMetrics

Preconditions

vmalert is installed. You can obtain it by building it from source, downloading it from the GitHub releases page, or using the docker image for the container ecosystem (such as docker, k8s, etc.).
Alertmanager is installed.
You have a single or cluster deployment in Managed VictoriaMetrics.
If you are using helm, add the VictoriaMetrics helm chart repository to your helm repositories. This step is optional.
If you are using vmoperator, make sure that it and its CRDs are installed. This step is also optional.

Setup

Alerting and recording rules file(s)

You need to prepare file(s) with alerting or recording rules.

An example file with one alerting rule.

alerts.yml

{% raw %}

groups:
  - name: common
    rules:
    - alert: instanceIsDown
      for: 1m
      expr: up == 0
      labels:
        severity: critical
      annotations:
        summary: "{{ $labels.job }} instance: {{$labels.instance }} is not up"
        description: "Job {{ $labels.job }}  instance: {{$labels.instance }} is not up for the last 1 minute"

{% endraw %}

Managed VictoriaMetrics access token and deployment endpoint

To use vmalert with Managed VictoriaMetrics, you must create a read/write token, or use an existing one. The token must have write access to ingest recording rules, ALERTS and ALERTS_FOR_STATE metrics, and read access for rules evaluation.

For instructions on how to create tokens, please refer to this section of the documentation.

Single-Node

Cluster

vmalert configuration

Single-Node

Binary

export TOKEN=81e8226e-****-****-****-************ 
export MANAGED_VM_URL=https://gw-c15-1c.cloud.victoriametrics.com
export ALERTMANAGER_URL=http://localhost:9093 
./vmalert -rule=alerts.yml -datasource.url=$MANAGED_VM_URL -datasource.bearerToken=$TOKEN -notifier.url=$ALERTMANAGER_URL -remoteWrite.url=$MANAGED_VM_URL -remoteWrite.bearerToken=$TOKEN -remoteRead.url=$MANAGED_VM_URL -remoteRead.bearerToken=$TOKEN

Docker

export TOKEN=81e8226e-****-****-****-************ 
export MANAGED_VM_URL=https://gw-c15-1c.cloud.victoriametrics.com
export ALERTMANAGER_URL=http://alertmanager:9093 
docker run -it -p 8080:8080 -v $(pwd)/alerts.yml:/etc/alerts/alerts.yml victoriametrics/vmalert:v1.87.1 -datasource.url=$MANAGED_VM_URL -datasource.bearerToken=$TOKEN -remoteRead.url=$MANAGED_VM_URL -remoteRead.bearerToken=$TOKEN  -remoteWrite.url=$MANAGED_VM_URL -remoteWrite.bearerToken=$TOKEN  -notifier.url=$ALERTMANAGER_URL -rule="/etc/alerts/*.yml"

Helm Chart

{% raw %}

export TOKEN=81e8226e-****-****-****-************
export MANAGED_VM_URL=https://gw-c15-1c.cloud.victoriametrics.com
export ALERTMANAGER=http://alertmanager:9093
cat <<EOF | helm install vmalert vm/victoria-metrics-alert -f -
server:
 datasource:
   url: $MANAGED_VM_URL
   bearer:
     token: $TOKEN
 remote:
   write:
     url: $MANAGED_VM_URL
     bearer:
       token: $TOKEN
   read:
     url: $MANAGED_VM_URL
     bearer:
       token: $TOKEN
 notifier:
   alertmanager:
     url: $ALERTMANAGER
 config:
   alerts:
     groups:
       - name: common
         rules:
           - alert: instanceIsDown
             for: 1m
             expr: up == 0
             labels:
               severity: critical
             annotations:
               summary: "{{ $labels.job }} instance: {{$labels.instance }} is not up"
               description: "Job {{ $labels.job }}  instance: {{$labels.instance }} is not up for the last 1 minute"
EOF

{% endraw %}

VMalert CRD for vmoperator

export TOKEN=81e8226e-****-****-****-************
export MANAGED_VM_URL=https://gw-c15-1c.cloud.victoriametrics.com
export ALERTMANAGER=http://alertmanager:9093
cat << EOF | kubectl apply -f -
apiVersion: operator.victoriametrics.com/v1beta1
kind: VMAlert
metadata:
 name: vmalert-managed-vm
spec:
 replicaCount: 1
 datasource:
   url: $MANAGED_VM_URL
   bearerTokenSecret:
     name: managed-token
     key: token
 remoteWrite:
   url: $MANAGED_VM_URL
   bearerTokenSecret:
     name: managed-token
     key: token
 remoteRead:
   url: $MANAGED_VM_URL
   bearerTokenSecret:
     name: managed-token
     key: token
 notifier:
   url: $ALERTMANAGER
 ruleSelector:
   matchLabels:
     type: managed
---
apiVersion: v1
kind: Secret
metadata:
 name: managed-token
stringData:
 token: $TOKEN
EOF

Testing

You can ingest metric that will raise an alert

export TOKEN=81e8226e-****-****-****-*************
export MANAGED_VM_URL=https://gw-c15-1c.cloud.victoriametrics.com/
curl -H "Authorization: Bearer $TOKEN" -X POST "$MANAGED_VM_URLapi/v1/import/prometheus" -d 'up{job="vmalert-test", instance="localhost"} 0'

Cluster

Binary

export TOKEN=76bc5470-****-****-****-************
export MANAGED_VM_READ_URL=https://gw-c15-1a.cloud.victoriametrics.com/select/0/prometheus/
export MANAGED_VM_WRITE_URL=https://gw-c15-1a.cloud.victoriametrics.com/insert/0/prometheus/
export ALERTMANAGER_URL=http://localhost:9093 
./vmalert -rule=alerts.yml -datasource.url=$MANAGED_VM_READ_URL -datasource.bearerToken=$TOKEN -notifier.url=$ALERTMANAGER_URL -remoteWrite.url=$MANAGED_VM_WRITE_URL -remoteWrite.bearerToken=$TOKEN -remoteRead.url=$MANAGED_VM_READ_URL -remoteRead.bearerToken=$TOKEN

Docker

export TOKEN=76bc5470-****-****-****-************
export MANAGED_VM_READ_URL=https://gw-c15-1a.cloud.victoriametrics.com/select/0/prometheus/
export MANAGED_VM_WRITE_URL=https://gw-c15-1a.cloud.victoriametrics.com/insert/0/prometheus/
export ALERTMANAGER_URL=http://alertmanager:9093 
docker run -it -p 8080:8080 -v $(pwd)/alerts.yml:/etc/alerts/alerts.yml victoriametrics/vmalert:v1.87.1 -datasource.url=$MANAGED_VM_READ_URL -datasource.bearerToken=$TOKEN -remoteRead.url=$MANAGED_VM_READ_URL -remoteRead.bearerToken=$TOKEN  -remoteWrite.url=$MANAGED_VM_WRITE_URL -remoteWrite.bearerToken=$TOKEN  -notifier.url=$ALERTMANAGER_URL -rule="/etc/alerts/*.yml"

Helm Chart

{% raw %}

export TOKEN=76bc5470-****-****-****-************
export MANAGED_VM_READ_URL=https://gw-c15-1a.cloud.victoriametrics.com/select/0/prometheus/
export MANAGED_VM_WRITE_URL=https://gw-c15-1a.cloud.victoriametrics.com/insert/0/prometheus/
export ALERTMANAGER=http://alertmanager:9093
cat <<EOF | helm install vmalert vm/victoria-metrics-alert -f -
server:
 datasource:
   url: $MANAGED_VM_READ_URL
   bearer:
     token: $TOKEN
 remote:
   write:
     url: $MANAGED_VM_WRITE_URL
     bearer:
       token: $TOKEN
   read:
     url: $MANAGED_VM_READ_URL
     bearer:
       token: $TOKEN
 notifier:
   alertmanager:
     url: $ALERTMANAGER
 config:
   alerts:
     groups:
       - name: common
         rules:
           - alert: instanceIsDown
             for: 1m
             expr: up == 0
             labels:
               severity: critical
             annotations:
               summary: "{{ $labels.job }} instance: {{$labels.instance }} is not up"
               description: "Job {{ $labels.job }}  instance: {{$labels.instance }} is not up for the last 1 minute"
EOF

{% endraw %}

VMalert CRD for vmoperator

export TOKEN=76bc5470-****-****-****-************
export MANAGED_VM_READ_URL=https://gw-c15-1a.cloud.victoriametrics.com/select/0/prometheus/
export MANAGED_VM_WRITE_URL=https://gw-c15-1a.cloud.victoriametrics.com/insert/0/prometheus/
export ALERTMANAGER=http://alertmanager:9093
cat << EOF | kubectl apply -f -
apiVersion: operator.victoriametrics.com/v1beta1
kind: VMAlert
metadata:
 name: vmalert-managed-vm
spec:
 replicaCount: 1
 datasource:
   url: $MANAGED_VM_READ_URL
   bearerTokenSecret:
     name: managed-token
     key: token
 remoteWrite:
   url: $MANAGED_VM_WRITE_URL
   bearerTokenSecret:
     name: managed-token
     key: token
 remoteRead:
   url: $MANAGED_VM_READ_URL
   bearerTokenSecret:
     name: managed-token
     key: token
 notifier:
   url: $ALERTMANAGER
 ruleSelector:
   matchLabels:
     type: managed
---
apiVersion: v1
kind: Secret
metadata:
 name: managed-token
stringData:
 token: $TOKEN
EOF

Testing

You can ingest metric that will raise an alert

export TOKEN=76bc5470-****-****-****-************
export MANAGED_VM_WRITE_URL=https://gw-c15-1a.cloud.victoriametrics.com/insert/0/prometheus/
curl -H "Authorization: Bearer $TOKEN" -X POST "$MANAGED_VM_WRITE_URLapi/v1/import/prometheus" -d 'up{job="vmalert-test", instance="localhost"} 0'

9.7 KiB Raw Blame History

Alerting with vmalert and Managed VictoriaMetrics

Preconditions

Setup

Alerting and recording rules file(s)

Managed VictoriaMetrics access token and deployment endpoint

Single-Node

Cluster

vmalert configuration

Single-Node

Binary

Docker

Helm Chart

VMalert CRD for vmoperator

Testing

Cluster

Binary

Docker

Helm Chart

VMalert CRD for vmoperator

Testing

9.7 KiB

Raw Blame History