OpenSearch Tamed for Production Reliability
Open-source alternative, lower lock-in, same operational challenges. We standardize indices, manage retention, and make it operationally manageable.
What we deliver with OpenSearch
Index and template standards
Component templates, explicit mappings, and index patterns that keep query behaviour stable as new data sources arrive.
Retention without surprises
ISM policies with tiering and deletion, sized against storage growth so the cluster does not fill up during a busy week.
Cluster sizing and upgrades
Node roles, shard allocation, and a tested upgrade path that accounts for plugin compatibility before the maintenance window.
Security and multi-tenancy
The security plugin configured with roles, backend role mapping, and tenant separation so dashboards do not leak between teams.
Dashboards and operations
Saved objects managed as code, snapshot verification, and runbooks for the failures that actually page someone.
How we work with OpenSearch
OpenSearch is operationally honest: it does what you configure and nothing more. We begin with templates and ISM, then validate snapshot restore, then work on query performance. Teams that reverse that order usually end up rebuilding indices they cannot easily reindex.
Because OpenSearch moves quickly, upgrades deserve their own plan. We check plugin and API compatibility against a staging cluster with production-like data, so the upgrade is a scheduled task rather than an incident.
OpenSearch in production: where it breaks
The same handful of problems show up on almost every engagement. Naming them early is cheaper than discovering them at 2am.
Plugins break on upgrade
A custom plugin without a compatible build stalls the whole rollout. We verify compatibility in staging before booking a maintenance window.
Shard over-allocation
Too many small shards per node degrade search and recovery. Target shard size plus ISM keeps the count sane.
Untested snapshots
Snapshots exist, restores fail. A quarterly restore into a scratch cluster proves the recovery path works.
When to use OpenSearch
Cost-sensitive search and logging where licence fees are a real constraint.
AWS-native estates already using managed OpenSearch domains.
Teams migrating from Elasticsearch who want to keep queries and dashboards working.