Skip to content

Latest commit

 

History

History
180 lines (142 loc) · 14 KB

migration.md

File metadata and controls

180 lines (142 loc) · 14 KB
Raw

Migration

This is a migration doc for the Observability for Kubernetes Operator from the manual and Helm Collector and proxy installation. If you want to test the new Operator in parallel with your existing manual or Helm installation, use the wavefront-allow-legacy-install.yaml template.

Table of Contents

Migrate from Helm Installation

1. Install the Operator

Follow the Operator installation instructions.

In your wavefront.yaml,

  • set spec.allowLegacyInstall to true
  • set spec.clusterName to <YOUR HELM CLUSTER NAME>-operator

2. Modify Your wavefront.yaml

Modify your wavefront.yaml to match your Helm installation based on the information below. The following table lists the mapping of configurable parameters of the Wavefront Helm chart to Observability for Kubernetes Operator Custom Resource. See Custom Resource Scenarios for examples or refer to wavefront.com_wavefronts.yaml for information on all Custom Resource fields.

Helm Collector Parameter Observability for Kubernetes Operator Custom Resource spec. Description
clusterName clusterName ClusterName is a unique name for the Kubernetes cluster to be identified via a metric tag in Operations for Applications.
wavefront.url wavefrontUrl The URL of your product cluster. Ex: https://<your_cluster>.wavefront.com.
wavefront.token wavefrontTokenSecret WavefrontTokenSecret is the name of the secret that contains a Wavefront API Token.
collector.enabled dataCollection.metrics.enable Metrics holds the configuration for node and cluster collectors.
collector.interval dataCollection.metrics.defaultCollectionInterval Default metrics collection interval. Defaults to 60s.
collector.useProxy NA The earlier Collector config parameter was meant to be used to support direct ingestion, which the Operator doesn't support.
collector.proxyAddress dataExport.externalWavefrontProxy.Url Url is the proxy URL that the collector sends metrics to.
collector.tags dataCollection.metrics.tags Map of tags (key/value) to add to all metrics collected.
collector.filters.metricDenyList dataCollection.metrics.filters.denyList List of metric patterns to deny.
collector.filters.metricAllowList dataCollection.metrics.filters.allowList List of metric patterns to allow.
collector.discovery.enabled dataCollection.metrics.enableDiscovery Rules based and Prometheus endpoints auto-discovery. Defaults to true.
collector.resources dataCollection.metrics.nodeCollector.resources dataCollection.metrics.clusterCollector.resources Compute resources required by the node and cluster collector containers.
proxy.enabled dataExport.wavefrontProxy.enable Enable is whether to enable the Wavefront proxy. Defaults to true. Disable to use dataExport.externalWavefrontProxy.Url.
proxy.port dataExport.wavefrontProxy.metricPort MetricPort is the port for sending Operations for Applications data format metrics. Defaults to 2878.
proxy.httpProxyHost dataExport.wavefrontProxy.httpProxy.secret Name of the secret containing the HttpProxy configuration.
proxy.httpProxyPort dataExport.wavefrontProxy.httpProxy.secret Name of the secret containing the HttpProxy configuration.
proxy.useHttpProxyCAcert dataExport.wavefrontProxy.httpProxy.secret Name of the secret containing the HttpProxy configuration.
proxy.httpProxyUser dataExport.wavefrontProxy.httpProxy.secret Name of the secret containing the HttpProxy configuration.
proxy.httpProxyPassword dataExport.wavefrontProxy.httpProxy.secret Name of the secret containing the HttpProxy configuration.
proxy.tracePort dataExport.wavefrontProxy.tracing.wavefront.port Port for sending distributed Operations for Applications format tracing data (usually 30000).
proxy.jaegerPort dataExport.wavefrontProxy.tracing.jaeger.port Port for Jaeger format tracing data (usually 30001).
proxy. traceJaegerHttpListenerPort dataExport.wavefrontProxy.tracing.jaeger. httpPort HttpPort for Jaeger Thrift format data (usually 30080).
proxy. traceJaegerGrpcListenerPort dataExport.wavefrontProxy.tracing.jaeger. grpcPort GrpcPort for Jaeger gRPC format data (usually 14250).
proxy.zipkinPort dataExport.wavefrontProxy.tracing.zipkin.port Port for Zipkin format tracing data (usually 9411).
proxy.traceSamplingRate dataExport.wavefrontProxy.tracing.wavefront. samplingRate SamplingRate Distributed tracing data sampling rate (0 to 1).
proxy.traceSamplingDuration dataExport.wavefrontProxy.tracing.wavefront. samplingDuration SamplingDuration When set to greater than 0, spans that exceed this duration will force trace to be sampled (ms).
proxy. traceJaegerApplicationName dataExport.wavefrontProxy.tracing.jaeger. applicationName Custom application name for traces received on Jaeger's HTTP or gRPC port.
proxy. traceZipkinApplicationName dataExport.wavefrontProxy.tracing.zipkin. applicationName Custom application name for traces received on Zipkin's port.
proxy.histogramPort dataExport.wavefrontProxy.histogram.port Port for Operations for Applications histogram distributions (usually 40000)
proxy.histogramMinutePort dataExport.wavefrontProxy.histogram.minutePort Port to accumulate 1-minute based histograms in the Operations for Applications data format (usually 40001).
proxy.histogramHourPort dataExport.wavefrontProxy.histogram.hourPort Port to accumulate 1-hour based histograms in the Operations for Applications data format (usually 40002).
proxy.histogramDayPort dataExport.wavefrontProxy.histogram.dayPort Port to accumulate 1-day based histograms in the Operations for Applications data format (usually 40002).
proxy.deltaCounterPort dataExport.wavefrontProxy.deltaCounterPort Port to send delta counters in the Operations for Applications data format (usually 50000).
proxy.args dataExport.wavefrontProxy.args Additional Wavefront proxy properties can be passed as command line arguments in the --<property_name> <value> format. Multiple properties can be specified.
proxy.preprocessor.rules.yaml dataExport.wavefrontProxy.preprocessor Name of the configmap containing a rules.yaml key with proxy preprocessing rules.

If you have a Collector configuration with parameters not covered above, please open an issue in this repository.

3. Re-apply Your wavefront.yaml File

kubectl apply -f <path_to_your_wavefront.yaml>

4. Verify The Operator Status

Check the status of your installation:

kubectl get wavefront -n observability-system

You should eventually see a table that looks something like the following:

NAME        STATUS    PROXY           CLUSTER-COLLECTOR   NODE-COLLECTOR   LOGGING        AGE    MESSAGE
wavefront   Healthy   Running (1/1)   Running (1/1)       Running (3/3)                   2m4s   All components are healthy

If STATUS is Unhealthy, check troubleshooting.

5. Verify That You Are Receiving Metrics from the Operator

  • Go to the Kubernetes Summary Dashboard, set the cluster to <YOUR HELM CLUSTER NAME>-operator, and verify that you are receiving metrics. It may take some time for metrics to populate this dashboard.
  • Verify that other important metrics are present with a cluster="<YOUR HELM CLUSTER NAME>-operator" tag.

6. Uninstall Helm

helm uninstall wavefront --namespace wavefront

7. Clean Up the wavefront.yaml File

In your wavefront.yaml,

  • (Optional) Update spec.clusterName to <YOUR HELM CLUSTER NAME>
  • Update spec.allowLegacyInstall to false
kubectl apply -f <path_to_your_wavefront.yaml>

Migrate from Manual Installation

Wavefront Proxy Configuration

References:

  • See Custom Resource Scenarios for proxy configuration examples.
  • Copy and save your existing Collector configMaps and any other configurations.
  • Uninstall your currently deployed Collector and Wavefront proxy.
  • Make sure the Observability for Kubernetes Operator is already installed.
  • Create a Kubernetes secret with your Wavefront API Token kubectl create -n observability-system secret generic wavefront-secret --from-literal token=YOUR_WAVEFRONT_TOKEN

Most of the proxy configurations could be set using environment variables for proxy container. Here are the different proxy environment variables and how they map to the Operator config.

Proxy Environment Variables Observability for Kubernetes Operator Custom Resource spec
WAVEFRONT_URL wavefrontUrl Ex: https://<your_cluster>.wavefront.com
WAVEFRONT_TOKEN WAVEFRONT_TOKEN is now stored in a Kubernetes secret; see References above.
WAVEFRONT_PROXY_ARGS dataExport.wavefrontProxy.* Refer to the below table for details.

Below are the proxy arguments that are specified in WAVEFRONT_PROXY_ARGS, which are currently supported natively in the Custom Resource.

Wavefront Proxy args Observability for Kubernetes Operator Custom Resource spec
--preprocessorConfigFile dataExport.wavefrontProxy.preprocessor ConfigMap
--proxyHost dataExport.wavefrontProxy.httpProxy.secret Secret
--proxyPort dataExport.wavefrontProxy.httpProxy.secret Secret
--proxyUser dataExport.wavefrontProxy.httpProxy.secret Secret
--proxyPassword dataExport.wavefrontProxy.httpProxy.secret Secret
--pushListenerPorts dataExport.wavefrontProxy.metricPort
--deltaCounterPorts dataExport.wavefrontProxy.deltaCounterPort
--traceListenerPorts dataExport.wavefrontProxy.tracing.wavefront.port
--traceSamplingRate dataExport.wavefrontProxy.tracing.wavefront.samplingRate
--traceSamplingDuration dataExport.wavefrontProxy.tracing.wavefront.samplingDuration
--traceZipkinListenerPorts dataExport.wavefrontProxy.tracing.zipkin.port
--traceZipkinApplicationName dataExport.wavefrontProxy.tracing.zipkin.applicationName
--traceJaegerListenerPorts dataExport.wavefrontProxy.tracing.jaeger.port
--traceJaegerHttpListenerPorts dataExport.wavefrontProxy.tracing.jaeger.httpPort
--traceJaegerGrpcListenerPorts dataExport.wavefrontProxy.tracing.jaeger.grpcPort
--traceJaegerApplicationName dataExport.wavefrontProxy.tracing.jaeger.applicationName
--histogramDistListenerPorts dataExport.wavefrontProxy.histogram.port
--histogramMinuteListenerPorts dataExport.wavefrontProxy.histogram.minutePort
--histogramHourListenerPorts dataExport.wavefrontProxy.histogram.hourPort
--histogramDayListenerPorts dataExport.wavefrontProxy.histogram.dayPort

Other supported Custom Resource configuration:

Wavefront Collector to Kubernetes Metrics Collector Configuration

Wavefront Collector ConfigMap changes:

  • Wavefront Collector ConfigMap changed from wavefront-collector to observability-system namespace.
  • sinks.proxyAddress changed from wavefront-proxy.default.svc.cluster.local:2878 to wavefront-proxy:2878.
  • Change collector.yaml to config.yaml

Custom Resource spec changes:

Other supported Custom Resource configurations:

  • dataCollection.metrics.filters: Used to set filtering to be applied to all the metrics collected by the Kubernetes Metrics Collector. See wavefront-collector-filtering.yaml for an example.
  • dataCollection.metrics.tolerations: Used to set pod tolerations to be applied to all data collection for Kubernetes Metrics Collector. See wavefront-daemonset-pod-tolerations.yaml for an example.
  • dataCollection.metrics.nodeCollector.resources: Used to set container resource request or limits for Kubernetes Metrics Collector (Node). See wavefront-pod-resources.yaml for an example.
  • dataCollection.metrics.clusterCollector.resources: Used to set container resource request or limits for Kubernetes Metrics Collector (Cluster). See wavefront-pod-resources.yaml for an example.

Future Support

If you have feedback, or come across something that cannot be configured with the Operator, please open an issue in this repository.