Link
Zebrium Docs

© Zebrium, INC. 2020

Search

Installation

Assumptions

  1. All pre-installation steps have been completed.
  2. Your Kubernetes cluster is running and has appropriate resources available as defined in the pre-installation sizing considerations.

STEP 1: Installing the Helm Chart

  1. Create a secret using the Harbor repository that was setup for you by Zebrium. Zebrium will provide you with your Harbor USERNAME and PASSWORD. 
    kubectl create secret docker-registry regcred --docker-server=harbor.ops.zebrium.com --docker-username=<USERNAME> --docker-password=<PASSWORD> --docker-email=<EMAIL> --namespace <NAMESPACE>
  2. Update your Helm chart override.yaml file with the secret from Step 1. 
    global.imagePullSecret.Name="<SECRET>"
  3. Add the Harbor repo to your Kubernetes cluster 
    helm repo add --username <USERNAME> --password <PASSWORD> <REPO_NAME> https://harbor.ops.zebrium.com/chartrepo/onprem
helm repo update
  4. Install the Zebrium On Prem Software 
    helm upgrade <RELEASE_NAME> -i --namespace <NAMESPACE> --version <VERSION> <REPO_NAME>/zebrium-onprem -f <override.yaml>

STEP 2: Configuring Your Account

Zebrium On Prem currently supports a single account where all data will be ingested. This account name was defined when your Helm chart was created. You must now create this same account name within your On Prem instance using the API.

Note: Substitute <COMPANY>, <DEPLOYMENT_NAME> and <FQHN> with those values defined in the Helm chart when completing the pre-installation steps.

curl -X POST --data '{ "action":"create_account","customer":"<COMPANY>","deployment_name":"<DEPLOYMENT_NAME>"}' http://<FQHN>/report/v1/deployment

The API will return a payload detailing the account information.

Note: the zapi_token field returned from the API needs to be saved. This token will be used to ingest data into Zebrium On Prem in Step 5. below.

STEP 3: Configuring Outbound Notifications

If you would like Root Cause report notifications to be sent to a destination other than the Slack webhook defined in your Helm chart, you can configure additional notification channels using the API. The full API definition is located here.

Supported outbound notification channels include:

  1. Slack
  2. Microsoft Teams
  3. Webhook (with the zebrium_incident payload)

Here is an example that creates a webhook notification channel that will send the zebrium_incident payload to an endpoint using Token-based authentication:

curl -X POST --data '{ "channel_type":"webhook", "customer":"<COMPANY>", "name":"incident webhook", "webhook":"http://99.999.99.99:8000/api/v1/incident", "auth_scheme":"token", "auth_header_prefix":"Token", "auth_token":"HDTSHSHDBSGSRERWJDJSODYDNDYDD" }' http://<FQHN>/report/v1/outbound_channel.create

Note: the payload of the zebrium_incident webhook is defined here.

STEP 4: Configuring AutoSupport (optional)

When you register for Zebrium On Prem, Zebrium will provide a secure auth token that can be used to send logs from the Zebrium On Prem software to Zebrium for remote monitoring (akin to Zebrium AutoSupport).

Note: Installing the Zebrium Kubernetes Log Collector will send logs from all namespaces in your Kubernetes cluster.

The log collector is deployed as a Kubernetes Helm chart as follows:

  1. kubectl create namespace zebrium
  2. helm install zlog-collector zlog-collector --namespace zebrium --repo https://raw.githubusercontent.com/zebrium/ze-kubernetes-collector/master/charts --set zebrium.collectorUrl=https://zapi03.zebrium.com,zebrium.authToken=AUTH_TOKEN_FROM_ZEBRIUM,zebrium.timezone=KUBERNETES_HOST_TIMEZONE

STEP 5: Ingesting data into your Zebrium On Prem instance

There are three supported methods to ingest data into Zebrium On Prem using your ZAPI Token and Endpoint:

  1. Zebrium CLI command
  2. Kubernetes Log Collector
  3. Elastic integration using Logstash

Obtaining your ZAPI Token and Endpoint

The ZAPI Token was provided in the response payload when creating your account. You can obtain your ZAPI Token anytime by using the List Deployments API here.

The ZAPI Endpoint IP address can be obtained using the following command:

kubectl [ -n <namespace> ] get service zapi

Failure Domain Boundary

Because Zebrium On Prem currently supports a single account with only one deployment, if you intend to ingest data from unrelated services/applications it is important to specify the ze_deployment_name label which essentially defines a failure domain boundary for anomaly correlation.

You will see in the examples provided below, how to specify the ze_deployment_name label for each of the three methods that can be used to ingest data.

Note: The ze_deployment_name must be a single word lowercase characters.

Using the CLI to ingest data

Instructions for downloading, configuring and using the Zebrium CLI are located here.

Here is an example that ingests a Jira log file into the atlassian failure domain (ze_deployment_name):

~/zapi/bin/ze up --file=jira.log --log=jira --ids=zid_host=jiraserver,ze_deployment_name=atlassian --auth=97453627rDGSDE67FDCA77BCE44 --url=http://34.72.193.228:443

Using the Kubernetes Log Collector to ingest data

If your application to be "monitored" is Kubernetes-based, this is the preferred method for sending logs to Zebrium On Prem.

The log collector is deployed as a Kubernetes Helm chart as follows:

  1. kubectl create namespace zebrium
  2. helm install zlog-collector zlog-collector --namespace zebrium --repo https://raw.githubusercontent.com/zebrium/ze-kubernetes-collector/master/charts --set zebrium.collectorUrl=http://ZAPI_ENDPOINT:443,zebrium.authToken=ZAPI_TOKEN,zebrium.deployment=ZE_DEPLOYMENT_NAME,zebrium.timezone=KUBERNETES_HOST_TIMEZONE

Note: Remember to substitute ZAPI_ENDPOINT, ZAPI_TOKEN and ZE_DEPLOYMENT_NAME

Using Logstash to ingest data

Instructions for configuring this integration are located here.

Note: Please contact Zebrium Support for assistance with configuration.