<#import "/templates/guide.adoc" as tmpl>
<#import "/templates/links.adoc" as links>

<@tmpl.guide
title="Deploy an AWS Lambda to guard against Split-Brain"
summary="Building block for loadbalancer resilience"
tileVisible="false" >

This {section} explains how to reduce the impact when split-brain scenarios occur between two sites in a multi-site deployment.

This deployment is intended to be used with the setup described in the <@links.ha id="concepts-multi-site"/> {section}.
Use this deployment with the other building blocks outlined in the <@links.ha id="bblocks-multi-site"/> {section}.

include::partials/blueprint-disclaimer.adoc[]

== Architecture
In the event of a network communication failure between the two sites in a multi-site deployment, it is no
longer possible for the two sites to continue to replicate data between themselves and the two sites
will become increasingly out-of-sync. As it is possible for subsequent Keycloak requests to be routed to different
sites, this may lead to unexpected behaviour as previous updates will not have been applied to both sites.

In such scenarios a quorum is commonly used to determine which sites are marked as online or offline, however as multi-site deployments only consist of two sites, this is not possible.
Instead, we leverage "`fencing`" to ensure that when one of the sites is unable to connect to the other site, only one site remains in the loadbalancer configuration and hence only this site is able to serve subsequent users requests.

Once the fencing procedure is triggered the replication between two {jdgserver_name} clusters in each site is no longer enabled and as a result the sites will be out-of-sync.
To recover from the out-of-sync state a manual re-sync is necessary as described in <@links.ha id="operate-synchronize" />.
This is why a site which is removed via fencing will not be re-added automatically when the network communication failure is resolved, but only after such a synchronisation using the manual procedure <@links.ha id="operate-site-online" />.

In this {section} we describe how to implement fencing using a combination of https://prometheus.io/docs/alerting/latest/overview/[Prometheus Alerts]
and AWS Lambda functions. A Prometheus Alert is triggered when split-brain is detected by the {jdgserver_name} server metrics,
which results in the Prometheus AlertManager calling the AWS Lambda based webhook. The triggered Lambda function inspects
the current Global Accelerator configuration and removes the site reported to be offline.

In a true split-brain scenario, where both sites are still up but network communication is down, it is possible that both
sites will trigger the webhook simultaneously. We guard against this by ensuring that only a single Lambda instance can be executed at
a given time.

== Prerequisites

* ROSA HCP based multi-site Keycloak deployment
* AWS CLI Installed
* AWS Global Accelerator loadbalancer
* `jq` tool installed

== Procedure
. Enable Openshift user alert routing
+
.Command:
[source,bash]
----
kubectl apply -f - << EOF
apiVersion: v1
kind: ConfigMap
metadata:
  name: user-workload-monitoring-config
  namespace: openshift-user-workload-monitoring
data:
  config.yaml: |
    alertmanager:
      enabled: true
      enableAlertmanagerConfig: true
EOF
kubectl -n openshift-user-workload-monitoring rollout status --watch statefulset.apps/alertmanager-user-workload
----
+
. [[aws-secret]]Decide upon a username/password combination which will be used to authenticate the Lambda webhook and create an AWS Secret storing the password
+
.Command:
[source,bash]
----
aws secretsmanager create-secret \
  --name webhook-password \ # <1>
  --secret-string changeme \ # <2>
  --region eu-west-1 # <3>
----
<1> The name of the secret
<2> The password to be used for authentication
<3> The AWS region that hosts the secret
+
. Create the Role used to execute the Lambda.
+
.Command:
[source,bash]
----
<#noparse>
FUNCTION_NAME= # <1>
ROLE_ARN=$(aws iam create-role \
  --role-name ${FUNCTION_NAME} \
  --assume-role-policy-document \
  '{
    "Version": "2012-10-17",
    "Statement": [
      {
        "Effect": "Allow",
        "Principal": {
          "Service": "lambda.amazonaws.com"
        },
        "Action": "sts:AssumeRole"
      }
    ]
  }' \
  --query 'Role.Arn' \
  --region eu-west-1 \ #<2>
  --output text
)
</#noparse>
----
<1> A name of your choice to associate with the Lambda and related resources
<2> The AWS Region hosting your Kubernetes clusters
+
. Create and attach the 'LambdaSecretManager' Policy so that the Lambda can access AWS Secrets
+
.Command:
[source,bash]
----
<#noparse>
POLICY_ARN=$(aws iam create-policy \
  --policy-name LambdaSecretManager \
  --policy-document \
  '{
      "Version": "2012-10-17",
      "Statement": [
          {
              "Effect": "Allow",
              "Action": [
                  "secretsmanager:GetSecretValue"
              ],
              "Resource": "*"
          }
      ]
  }' \
  --query 'Policy.Arn' \
  --output text
)
aws iam attach-role-policy \
  --role-name ${FUNCTION_NAME} \
  --policy-arn ${POLICY_ARN}
</#noparse>
----
+
. Attach the `ElasticLoadBalancingReadOnly` policy so that the Lambda can query the provisioned Network Load Balancers
+
.Command:
[source,bash]
----
<#noparse>
aws iam attach-role-policy \
  --role-name ${FUNCTION_NAME} \
  --policy-arn arn:aws:iam::aws:policy/ElasticLoadBalancingReadOnly
</#noparse>
----
+
. Attach the `GlobalAcceleratorFullAccess` policy so that the Lambda can update the Global Accelerator EndpointGroup
+
.Command:
[source,bash]
----
<#noparse>
aws iam attach-role-policy \
  --role-name ${FUNCTION_NAME} \
  --policy-arn arn:aws:iam::aws:policy/GlobalAcceleratorFullAccess
</#noparse>
----
+
. Create a Lambda ZIP file containing the required fencing logic
+
.Command:
[source,bash]
----
<#noparse>
LAMBDA_ZIP=/tmp/lambda.zip
cat << EOF > /tmp/lambda.py

include::examples/generated/fencing_lambda.py[]

EOF
zip -FS --junk-paths ${LAMBDA_ZIP} /tmp/lambda.py
</#noparse>
----
+
. Create the Lambda function.
+
.Command:
[source,bash]
----
<#noparse>
aws lambda create-function \
  --function-name ${FUNCTION_NAME} \
  --zip-file fileb://${LAMBDA_ZIP} \
  --handler lambda.handler \
  --runtime python3.12 \
  --role ${ROLE_ARN} \
  --region eu-west-1 #<1>
</#noparse>
----
<1> The AWS Region hosting your Kubernetes clusters
+
. Expose a Function URL so the Lambda can be triggered as webhook
+
.Command:
[source,bash]
----
<#noparse>
aws lambda create-function-url-config \
  --function-name ${FUNCTION_NAME} \
  --auth-type NONE \
  --region eu-west-1 #<1>
</#noparse>
----
<1> The AWS Region hosting your Kubernetes clusters
+
. Allow public invocations of the Function URL
+
.Command:
[source,bash]
----
<#noparse>
aws lambda add-permission \
  --action "lambda:InvokeFunctionUrl" \
  --function-name ${FUNCTION_NAME} \
  --principal "*" \
  --statement-id FunctionURLAllowPublicAccess \
  --function-url-auth-type NONE \
  --region eu-west-1 # <1>
</#noparse>
----
<1> The AWS Region hosting your Kubernetes clusters
+
. Configure the Lambda's Environment variables:
+
.. In each Kubernetes cluster, retrieve the exposed {jdgserver_name} URL endpoint:
+
[source,bash]
----
<#noparse>
kubectl -n ${NAMESPACE} get route infinispan-external -o jsonpath='{.status.ingress[].host}' # <1>
</#noparse>
----
<1> Replace `$\{NAMESPACE}` with the namespace containing your {jdgserver_name} server
+
.. Upload the desired Environment variables
+
[source,bash]
----
<#noparse>
ACCELERATOR_NAME= # <1>
LAMBDA_REGION= # <2>
CLUSTER_1_NAME= # <3>
CLUSTER_1_ISPN_ENDPOINT= # <4>
CLUSTER_2_NAME= # <5>
CLUSTER_2_ISPN_ENDPOINT= # <6>
INFINISPAN_USER= # <7>
INFINISPAN_USER_SECRET= # <8>
WEBHOOK_USER= # <9>
WEBHOOK_USER_SECRET= # <10>

INFINISPAN_SITE_ENDPOINTS=$(echo "{\"${CLUSTER_NAME_1}\":\"${CLUSTER_1_ISPN_ENDPOINT}\",\"${CLUSTER_2_NAME}\":\"${CLUSTER_2_ISPN_ENDPOINT\"}" | jq tostring)
aws lambda update-function-configuration \
    --function-name ${ACCELERATOR_NAME} \
    --region ${LAMBDA_REGION} \
    --environment "{
      \"Variables\": {
        \"INFINISPAN_USER\" : \"${INFINISPAN_USER}\",
        \"INFINISPAN_USER_SECRET\" : \"${INFINISPAN_USER_SECRET}\",
        \"INFINISPAN_SITE_ENDPOINTS\" : ${INFINISPAN_SITE_ENDPOINTS},
        \"WEBHOOK_USER\" : \"${WEBHOOK_USER}\",
        \"WEBHOOK_USER_SECRET\" : \"${WEBHOOK_USER_SECERT}\",
        \"SECRETS_REGION\" : \"eu-central-1\"
      }
    }"
</#noparse>
----
+
<1> The name of the AWS Global Accelerator used by your deployment
<2> The AWS Region hosting your Kubernetes cluster and Lambda function
<3> The name of one of your {jdgserver_name} sites as defined in <@links.ha id="deploy-infinispan-kubernetes-crossdc" />
<4> The {jdgserver_name} endpoint URL associated with the CLUSER_1_NAME site
<5> The name of the second {jdgserver_name} site
<6> The {jdgserver_name} endpoint URL associated with the CLUSER_2_NAME site
<7> The username of a {jdgserver_name} user which has sufficient privileges to perform REST requests on the server
<8> The name of the AWS secret containing the password associated with the {jdgserver_name} user
<9> The username used to authenticate requests to the Lambda Function
<10> The name of the AWS secret containing the password used to authenticate requests to the Lambda function
+
. Retrieve the Lambda Function URL
+
.Command:
[source,bash]
----
<#noparse>
aws lambda get-function-url-config \
  --function-name ${FUNCTION_NAME} \
  --query "FunctionUrl" \
  --region eu-west-1 \#<1>
  --output text
</#noparse>
----
<1> The AWS region where the Lambda was created

+
.Output:
[source,bash]
----
https://tjqr2vgc664b6noj6vugprakoq0oausj.lambda-url.eu-west-1.on.aws
----
. In each Kubernetes cluster, configure a Prometheus Alert routing to trigger the Lambda on split-brain
+
.Command:
[source,bash]
----
<#noparse>
NAMESPACE= # The namespace containing your deployments
kubectl apply -n ${NAMESPACE} -f - << EOF
include::examples/generated/ispn-site-a.yaml[tag=fencing-secret]
</#noparse>
---
include::examples/generated/ispn-site-a.yaml[tag=fencing-alert-manager-config]
---
include::examples/generated/ispn-site-a.yaml[tag=fencing-prometheus-rule]
----
<1> The username required to authenticate Lambda requests
<2> The password required to authenticate Lambda requests
<3> The Lambda Function URL
<4> The namespace value should be the namespace hosting the Infinispan CR and the site should be the remote site defined
by `spec.service.sites.locations[0].name` in your Infinispan CR
<5> The name of your local site defined by `spec.service.sites.local.name` in your Infinispan CR
<6> The DNS of your Global Accelerator

== Verify

To test that the Prometheus alert triggers the webhook as expected, perform the following steps to simulate a split-brain:

. In each of your clusters execute the following:
+
.Command:
[source,bash]
----
<#noparse>
kubectl -n openshift-operators scale --replicas=0 deployment/infinispan-operator-controller-manager #<1>
kubectl -n openshift-operators rollout status -w deployment/infinispan-operator-controller-manager
kubectl -n ${NAMESPACE} scale --replicas=0 deployment/infinispan-router #<2>
kubectl -n ${NAMESPACE} rollout status -w deployment/infinispan-router
</#noparse>
----
<1> Scale down the {jdgserver_name} Operator so that the next step does not result in the deployment being recreated by the operator
<2> Scale down the Gossip Router deployment.Replace `$\{NAMESPACE}` with the namespace containing your {jdgserver_name} server
+
. Verify the `SiteOffline` event has been fired on a cluster by inspecting the *Observe* -> *Alerting* menu in the Openshift
console
+
. Inspect the Global Accelerator EndpointGroup in the AWS console and there should only be a single endpoint present
+
. Scale up the {jdgserver_name} Operator and Gossip Router to re-establish a connection between sites:
+
.Command:
[source,bash]
----
<#noparse>
kubectl -n openshift-operators scale --replicas=1 deployment/infinispan-operator-controller-manager
kubectl -n openshift-operators rollout status -w deployment/infinispan-operator-controller-manager
kubectl -n ${NAMESPACE} scale --replicas=1 deployment/infinispan-router #<1>
kubectl -n ${NAMESPACE} rollout status -w deployment/infinispan-router
</#noparse>
----
<1> Replace `$\{NAMESPACE}` with the namespace containing your {jdgserver_name} server
+
. Inspect the `vendor_jgroups_site_view_status` metric in each site. A value of `1` indicates that the site is reachable.
+
. Update the Accelerator EndpointGroup to contain both Endpoints. See the <@links.ha id="operate-site-online" /> {section} for details.

== Further reading

* <@links.ha id="operate-site-online" />
* <@links.ha id="operate-site-offline" />

</@tmpl.guide>