Skip to content

Latest commit

 

History

History
64 lines (43 loc) · 2.89 KB

monitoring-with-prometheus.adoc

File metadata and controls

64 lines (43 loc) · 2.89 KB

Monitoring with Prometheus

Table of Contents

The main aim in this exercise is to expose the Data Virtualization servie metrics that was deployed in the previous steps using Prometheus.

Working with existing Prometheus Instance

the Teiid Operator will automatically find the Prometheus instance in the cluster based on the matching labels configured as "team=middleware", which are default labels that Teiid Operator is configured to scan with.

serviceMonitorSelector:
  matchLabels:
    team: middleware

If Teiid Operator is needs to scan on different matching labels, configure a environment property "PROMETHEUS_MONITOR_LABEL_KEY" and "PROMETHEUS_MONITOR_LABEL_VALUE". For example

PROMETHEUS_MONITOR_LABEL_KEY=foo
PROMETHEUS_MONITOR_LABEL_VALUE=bar

on the operator deployment environment properties.

Note
 The operator automatically adds the configuration for metrics that it going to expose for given virtualization. However, if user wants expose different set of metrics they can provide a config-map with name teiid-prometheus-config with key called rules and then provide the full rules in yaml content in there. During the deployment if Operator finds existence of this config-map then it will use it for prometheus configuration instead of default one.

Prometheus Operator

Note that in this example, the installation of Prometheus is done using using a Prometheus Operator.

Note
 Installing the Operator and the Prometheus instance is only for testing purposes. In production scenarios there should be common Prometheus installed for set of applications that every service uses.

Install

TODO - start with https://github.com/coreos/prometheus-operator there are several options including a helm chart.

For convenience the Operator install YAML file included here, which will install the Prometheus Operator in myproject namespace. In order to install the user needs to have cluster admin permissions.

# login as admin user, or user with cluster admin permissions
oc login

# install the operator in myproject namespace (edit to place in different namespace)
oc create -f prometheus-operator.yaml

# create a Prometheus server in "myproject" namespace (edit to place in different namespace)
oc create -f prometheus-server.yaml

Once the above is installed, the Teiid Operator will automatically find the Prometheus instance based on the matching labels configured "team=dv", which are default labels that Teiid Operator is configured to scan with.

serviceMonitorSelector:
  matchLabels:
    team: dv

Once the Prometheus instance is found, Teiid Operator automatically configures the ServiceMonitor to capture the metrics.

Graphana

User once the Prometheus instance is configured can also install a Graphana dashboard to monitor the Dashboard of metrics. Installing Graphana and configuring of dashboard is currently left as user exercise.