A Fortune 500 Tech Firm Future-Proofs Elasticsearch with Vendor Helm, ILM, and Cross-Cluster Resilience
Client
The client is a globally recognized semiconductor and enterprise technology company operating high-volume Elasticsearch workloads across multiple Kubernetes clusters. With over 21 TB of data per region, their infrastructure required modernization to ensure resilience, observability, and scalability.
Project Context
The client was running Elasticsearch on Elastic Cloud on Kubernetes (ECK) v2.5 using custom Helm charts. They sought to upgrade to ECK v2.14 using official Helm support, improve observability, and strengthen configuration standards across global environments.
Challenges
- Complex custom Helm configuration
- Cross-region deployment and zero-downtime requirement
- Performance tuning for 128GB nodes
- Resilience features like rack awareness, forced awareness, and CCR
- Troubleshooting Kubernetes observability components
Solution
- Planned and tested ECK upgrade path with snapshot validation
- Customized Helm values.yaml for node roles, JVM tuning, and memory
- Benchmarked performance using Elasticsearch Rally
- Implemented ILM hot-warm-cold, CCS, CCR, and rack awareness
- Enabled Filebeat, APM, and Fleet for production observability
Project Objectives
- Migrate to ECK 2.14 with zero downtime and no data loss
- Transition from custom Helm charts to vendor-maintained Helm
- Improve resource efficiency for large-memory nodes (up to 128GB)
- Implement ILM, CCS/CCR, and rack awareness
- Enable full-stack observability in Kubernetes
Solution Delivery
SquareShift executed a phased migration strategy with snapshot-based safety, transitioned to Elastic’s official Helm charts, and tuned deployment performance using Rally. They also strengthened observability using Elastic APM and Filebeat within Kubernetes environments.
Testimonial
SquareShift helped us scale our global Elasticsearch operations with modern tooling, resilience, and zero downtime.