/
Post-Campaign Infrastructure Optimization

Post-Campaign Infrastructure Optimization

Owned by Nikhil Mulinti
Last updated: Nov 05, 2024
3 min read

 

Scaling Down Deployments

scale_down.sh Script

The scale_down.sh script is designed to scale down specified core services within the egov namespace to a single replica and scale down all other services to zero replicas. It also records the original replica counts for future scaling operations.

#!/bin/bash # File to store the replica count for each deployment replica_info_file="replica_counts.txt" > "$replica_info_file" # Clear the file before starting # List of core services to be maintained and scaled down to 1 replica core_services=( "egov-accesscontrol" "egov-enc-service" "egov-localization" "egov-location" "egov-mdms-service" "egov-user" "zuul" "dss-service" "boundary-service" ) # Function to save and scale down replicas scale_down() { service=$1 namespace=$2 current_replicas=$(kubectl get deployment "$service" -n "$namespace" -o jsonpath='{.spec.replicas}') echo "$service $current_replicas" >> "$replica_info_file" kubectl scale deployment "$service" --replicas=1 -n "$namespace" sleep 1 echo "$service scaled down." } # Scale down core services in egov namespace to 1 replica for service in "${core_services[@]}"; do echo "Scaling down deployment $service in namespace egov to 1 replica..." scale_down "$service" "egov" done # Scale Kibana in the es-cluster-v8 namespace scale_down "release-name-kibana" "es-cluster-v8" # Scale down all other services in the egov namespace to 0 replicas all_services=$(kubectl get deployments -n egov -o jsonpath='{.items[*].metadata.name}') for service in $all_services; do if [[ ! " ${core_services[@]} " =~ " $service " ]]; then echo "Scaling down deployment $service in namespace egov to 0 replicas..." scale_down "$service" "egov" kubectl scale deployment "$service" --replicas=0 -n egov fi done

Execution Steps

  1. Ensure Executable Permissions:

    chmod +x scale_down.sh

  2. Run the Script:

    ./scale_down.sh

Scaling Up Deployments

scaleup.sh Script

The scaleup.sh script restores the replica counts of deployments based on the information saved in replica_counts.txt.

#!/bin/bash # File with saved replica counts replica_info_file="replica_counts.txt" # Check if the replica info file exists if [[ ! -f "$replica_info_file" ]]; then echo "Replica info file not found! Exiting." exit 1 fi # Read the file and scale each service to its original replica count while read -r line; do service=$(echo "$line" | awk '{print $1}') replicas=$(echo "$line" | awk '{print $2}') namespace="egov" echo "Scaling up deployment $service in namespace $namespace to $replicas replicas..." kubectl scale deployment "$service" --replicas="$replicas" -n "$namespace" sleep 1 echo "$service scaled up to $replicas replicas." done < "$replica_info_file"

Execution Steps

  1. Ensure Executable Permissions:

    chmod +x scaleup.sh

  2. Run the Script:

    ./scaleup.sh

Postgres RDS Termination and Deployment on Kubernetes

Backup and Terminate Existing RDS

  1. Take a Backup of the Existing Postgres RDS:

    • Use AWS RDS console or CLI to create a final snapshot.

    • Ensure the snapshot is successfully created before proceeding.

  2. Terminate the Existing Postgres RDS:

    • Navigate to the RDS console.

    • Select the Postgres RDS instance.

    • Choose Delete and follow the prompts to terminate the instance.

    Note: Ensure all necessary data is backed up and verified before termination.

Deploy Postgres on Kubernetes Pod

  1. Obtain Existing Database Credentials:

    • Get the Database Name:

      kubectl get cm egov-config -n egov -o jsonpath="{.data.db-name}"

      Example Output:

      mozhealthuat

    • Get the Database Username:

      kubectl get secret db -n egov -o jsonpath="{.data.username}" | base64 --decode

      Example Output:

      mozhealthuat

    • Get the Database Password:

      kubectl get secret db -n egov -o jsonpath="{.data.password}" | base64 --decode

      Example Output:

      [Your_DB_Password]

  2. Update the PostgreSQL Helm Chart and Deploy:

    • Retrieve PostgreSQL RDS Version from AWS:

      • Navigate to the AWS RDS console.

      • Locate the Postgres RDS instance and note the version (e.g., 12).

    • Install PostgreSQL Using Helm:

      helm install postgresql bitnami/postgresql \\ --set image.tag=12 \\ --set auth.username=mozhealthuat \\ --set auth.password=mozhealthuat \\ --set auth.database=mozhealthuat \\ --namespace egov

    Note: Adjust the image.tag, auth.username, auth.password, and auth.database values as per your requirements.

Post Deployment Configuration

Update ConfigMap

After deploying PostgreSQL, update the egov-config ConfigMap with the new database host and URL.

kubectl patch configmap egov-config -n egov --type='merge' -p '{ "data": { "db-host": "postgresql.egov.svc.cluster.local", "db-url": "jdbc:postgresql://postgresql.egov.svc.cluster.local:5432/mozhealthuat" } }'

Restart Deployments

To apply the new configurations, restart all deployments in the egov namespace.

kubectl get deployments -n egov -o name | xargs -n1 kubectl rollout restart -n egov

Related content