A Kubernetes Operator based on the Operator SDK to centralize the setup of 3rd party prometheus exporters on Kubernetes/OpenShift, with a collection of grafana dashboards.
By just providing a few parameters like dbHost or dbPort (operator manage the container image, port, argument, command, volumes... and also prometheus ServiceMonitor
and GrafanaDashboard
k8s objects), you can setup different prometheus exporters to monitor:
- The internals from different databases
- HTTP/TCP endpoints (availability, latency, SSL/TLS certificate expiration...)
- Any available cloudwatch metric from any AWS Service
- Sendgrid email statistics (delivered, bounces, errors. spam...)
Current prometheus exporters types
supported, managed by same prometheus-exporter-operator:
- memcached
- redis
- mysql
- postgresql
- sphinx
- manticore
- es (elasticsearch)
- cloudwatch
- probe (blackbox)
- sendgrid
The operator manages the lifecycle of the following objects:
- Deployment (one per CR)
- Service (one per CR)
- ServiceMonitor (optional, one per CR)
- GrafanaDashboard (optional, one per Namespace)
NOTE
Some exporters need some extra objects to be previously manually created in order to work (manual objects names need to be specified on required CR fields). This extra needed objects includes Secrets (credentials) or Configmaps (configuration files) on specific formats. Examples to help you create these extra objects are provided on examples directory for all exporter types.
If you modify the content of these extra needed objects (Secrets/Configmaps), exporters won't load them automatically, so you need to force a new pod creation by for example deleting the running pod.
Operator is available at OperatorHub.io (on both Kubernetes/OpenShift OLM catalogs)
- prometheus-operator v0.17.0+
- grafana-operator v3.0.0+
GrafanaDashboards
management is included in the operator:
- For each CR, a
GrafanaDashboard
(optional, enabled by defaultgrafanaDashboard.enabled: true
) is created, but actually operator manages a single dashboard type per Namespace (not per CR) - If you deploy for example different redis CRs, and you want to have the redis dashboard created, you need to enabled it on every redis CR with the same grafana-operator label selector (but actually, operator will just manage a single dashboard per Namespace shared accross all CRs from the same type)
- You can deploy the prometheus-exporter-operator with different operator versions on different Namespaces, so operator will create separate dashboards per Namespace (they won't collision, that's why dashboard name includes the Namespace)
- All grafana dashboards are preconfigured to use
CR_NAME
as the filter of all possible dashboards of every type (for examplestaging-system-memcached
) - In the future it is possible that
GrafanaDashboard
management get its own CRD separate fromPrometheusExporter
CRD (so you could have N PrometheusExporter CRs, and also an additonal single Dashboard CR per exporter type
PrometheusRules
management is NOT included in the operator (at least by the moment), because it depends on:
- What you need to monitor (maybe ones just need basic cpu/mem alerts, while others may be interested on specific alerts checking internals of a database)
- Why you want to be paged (severity warning/critical, minutes duration before firing an alert...)
- Customizable thresholds definition (it is something that depends on infrastructure dimensions...)
However, some examples of prometheus rules can be found at prometheus-rules directory.
You can contribute by:
- Raising any issues you find using Prometheus Exporter Operator
- Fixing issues by opening Pull Requests
- Submitting a patch or opening a PR
- Improving documentation
- Talking about Prometheus Exporter Operator
All bugs, tasks or enhancements are tracked as GitHub issues.
Prometheus Exporter Operator is under Apache 2.0 license. See the LICENSE file for details.