An Instana administrator is planning a major version upgrade for their self-hosted deployment. To minimize compatibility issues and service disruption, what is the recommended sequence for upgrading Instana components?

Instana agents are designed to be backward compatible with newer backend versions, so the recommended practice is to upgrade the backend first, then progressively update agents in waves. This approach ensures the backend is on the new version before agents begin reporting with any new data formats, and allows agents to continue reporting without interruption throughout the rollout. Simultaneous upgrades of all components increase risk and significantly complicate troubleshooting if issues arise. The correct answer is 'Upgrade the Instana backend (server and datastores) first, then progressively roll out updated agents across the environment'. The incorrect options — 'Upgrade all Instana agents first to ensure they support the new backend API before any backend components are changed', 'Upgrade all components simultaneously during a single maintenance window to avoid any version mismatch period', 'Upgrade Kafka and Zookeeper first, then Cassandra, then Elasticsearch, and finally the Instana server last' — are wrong because they do not align with IBM Instana's architecture or recommended practices for this scenario. Understanding this concept is essential for the Domain 5: Planning domain of the IBM Instana Observability certification.