Camel Monitor Operator 0.2.0 release
We’re pleased to announce the availability of Camel Monitor operator version 0.2.0. We decided to rename the old Camel Dashboard Operator in order to better clarify the intent of this tool. While the dashboard is something we’re still providing (via plugin UI), we thought that renaming the operator as more generically Camel Monitor was the right thing to do.
We want to give you the maximum flexibility and let use only the operator to perform any monitoring activity and leverage the UI plugin for dashboards.
CRD renaming
With the operator rename we thought we had to rename the custom resource we use as well. From now on we will refer to CamelMonitor (shortname, cmon) custom resource. The logic hasn’t changed. Whenever an operator discover a Camel application, it will create the related CamelMonitor. In this release we’ve added a few more metrics as illustrated in some following chapter.
Support CronJob
From this version onward the operator will be able to monitor also Camel applications deployed as CronJob.
Support KnativeService
Beside Deployment and CronJob we have also enabled the possibility to monitor any other custom resource that is finally producing a Deployment object baking a Camel application. It’s the case of Knative Service. So, even if you deal with any third party custom resource to manage Camel applications, if the final result is a Deployment, then, the Camel Monitor operator will be able to discover and manage it.
Camel K operator integration
Camel K is the best way to deploy Camel cloud natives applications. We are planning to make the Camel K operator and the Camel Monitor operator more integrated in order to provide a great experience both for deploying and for monitoring a Camel application in the cloud. In version 2.10 of Camel K it was introduced a mechanism to simplify the discovery of applications deployed via Camel K.
If you have both Camel K and Camel Monitor running in your cloud, release a Camel K Integration and you will get automatically a CamelMonitor with all the metrics you need to monitor your Integration.
Added new metrics
We have introduced a few more resource scraping. This is the completed list of the new available information:
- ProcessCPUUsed: How much CPU (in millicores) the process is consuming
- ProcessCPUMax: How much CPU (in millicores) the process is allowed to consume
- JVMMemoryUsed: How much memory (in bytes) the process is consuming
- JVMMemoryMax: How much memory (in bytes) the process is allowed to consume
- HasMemoryPressure: If true indicates that the memory usage is approximating dangerously to the cap
- HasCPUPressure: If true indicates that the CPU usage is approximating dangerously to the cap
The last two flags are a great way to immediately check if a giving application is suffering some kind of resource pressure.
Support Prometheus stack
In this release we have dedicated to support more strictly the Prometheus stack: Prometheus, Grafana and AlertManager.
The presence of camel-observability-services leverages Micrometer Prometheus technology. The exposure of the metrics endpoint enables the possibility for the operator to include the custom resource expected by Prometheus stack automatically.
If you have all the stack available in the cluster, you will be able to get automatically all custom resources required by the Prometheus stack to provide a deep level of monitoring and alerting.
PodMonitor
The operator is able to detect the presence of the Prometheus operator on the cluster, and, by default, include a PodMonitor associated with the Camel application monitored by the dashboard. Such a PodMonitor can be eventually used by the final user for more advanced monitoring activities.
You can setup the environment variable CREATE_PROMETHEUS_POD_MONITOR (by default it is enabled). It must be true to enable the creation of the Prometheus custom resource. Remove the variable or set to any other value to disable the feature.
You should also set the environment variable PROMETHEUS_LABEL in order to configure all the PodMonitor with the proper label selector expected by your existing Prometheus instance. The environment variable expect a single label formatted as key=value. By default, the operator will configure a label as camel.apache.org/prometheus=camel-monitor-operator. If you therefore want to create a dedicated instance to monitor all Camel Dashboard applications, you can configure your Prometheus to select such a default label.
GrafanaDashboards
The operator is also in charge to create a GrafanaDashboard custom resource equipped with a series of opinionated metrics for each Camel application it discovers. The result is an automatic dashboard that can be browsed in Grafana monitoring tool (which has to be available in the cluster).
NOTE: this feature is experimental.
When the operator detects the exposure of the Prometheus metrics endpoint and the presence of an existing Grafana operator (via the availability of GrafanaDashboard custom resource), it will create a new dashboard with the same name of the application. This will be picked by the Grafana instance and exposed in its interface.
You can setup the environment variable CREATE_GRAFANA_MONITOR (by default it is enabled). It must be true to enable the creation of the GrafanaDashboard custom resource. Remove the variable or set to any other value to disable the feature. Every dashboard created needs to setup the datasource to use: we control this configuration via the environment variable GRAFANA_DS which defaults to the value prometheus. If your Prometheus datasource has a different name, you’ll need to change the variable accordingly.
You should also set the environment variable GRAFANA_LABEL in order to configure all the dashboards created with the proper label selector expected by your existing Grafana instance. The environment variable expect a single label formatted as key=value. By default, the operator will configure a label as camel.apache.org/grafana=camel-monitor-operator. If you therefore want to create a dedicated instance to monitor all Camel Dashboard applications, you can configure your Grafana to select such a default label.
PrometheusRule
The operator is also in charge to create a single PrometheusRule custom resource equipped with a series of opinionated Camel alerts that could be used to integrate in any SRE system via Prometheus Alert Manager.
When the operator detects the existence of the Prometheus custom resources, it will create an alert named camel-dashboard-alerts in the same namespace where the operator is installed (the process is the same regardless if the operator is global or namespace scoped).
You can setup the environment variable CREATE_PROMETHEUS_RULE (by default it is enabled). It must be true to enable the creation of the PrometheusRule custom resource. Remove the variable or set to any other value to disable the feature.
You should also set the environment variable PROMETHEUS_RULE_LABEL in order to configure the alert rules created with the proper label selector expected by your existing Prometheus instance (i.e., via .spec.ruleSelector.matchLabels). The environment variable expect a single label formatted as key=value. By default, the operator will configure a label as camel.apache.org/alerts=camel-monitor-operator if none is specified.