Skip to content

Prepare for upgrades

Elastic Stack and managed-platform upgrades are coordinated through iVedha.

Do not upgrade Elasticsearch, Kibana, Kubernetes, or other managed infrastructure directly.

What iVedha manages

iVedha manages the supported platform upgrade process, including:

  • Elasticsearch and Kibana platform changes;
  • managed Kubernetes and supporting component changes when required;
  • upgrade sequencing;
  • platform health checks;
  • supported recovery procedures.

Customers are responsible for validating that their applications, integrations, and data continue to behave as expected after the upgrade.

Before an upgrade

Review the workloads that depend on the platform.

Check:

  • application Elasticsearch clients;
  • Elastic Agent and Fleet integrations;
  • OpenTelemetry instrumentation and integrations;
  • existing Logstash pipelines;
  • custom Elasticsearch plugins or extensions when used;
  • Kibana dashboards and saved objects;
  • SSO and authentication integrations;
  • application indexing and search behavior.

Review known compatibility or breaking changes that can affect these components.

Elastic Agent and Fleet

Before an Elastic Stack upgrade, review your Fleet-managed agents and integrations.

Confirm:

  • agents are reporting Healthy;
  • required integrations support the target platform version;
  • important policies are working normally;
  • no major agent or ingestion issue is already active.

Elastic Agent versions can be managed separately from the Elasticsearch platform version when supported.

Do not upgrade all agents simply because the Elasticsearch platform is being upgraded unless the target compatibility requires it.

OpenTelemetry workloads

For OpenTelemetry-based observability workloads, verify that:

  • applications continue to export OTLP successfully;
  • Elastic Agent OpenTelemetry integrations are healthy;
  • expected OTel-native data streams continue to receive data;
  • important service.name and resource attributes remain available.

Where integration packages have minimum Kibana or Elastic version requirements, review them before enabling new package versions.

Plan the maintenance window

Platform upgrades are normally coordinated using the configured maintenance window.

Before the upgrade:

  • notify affected application owners and users;
  • avoid unrelated major platform changes;
  • confirm critical applications can retry temporary connection failures;
  • identify the customer contacts responsible for post-upgrade validation.

Data protection

iVedha determines the appropriate platform recovery protection for the supported upgrade process.

Customers do not need to manually create snapshots for every managed upgrade.

For business-critical workloads, confirm that recovery requirements are understood before the change.

See back up and restore.

Validate after the upgrade

After iVedha confirms the platform upgrade has completed, validate the workloads that matter to your organization.

Check:

  • Kibana loads normally;
  • authentication and SSO work;
  • important users and role mappings still behave correctly;
  • Elastic Agents return to Healthy;
  • OpenTelemetry logs, metrics, and traces continue to arrive;
  • application indexing succeeds;
  • important search queries return expected results;
  • critical dashboards and visualizations work;
  • expected data streams and indices remain available.

Check the platform version

Users with appropriate permissions can review Elasticsearch version information through the Elasticsearch API.

For example:

curl --fail-with-body --user "<user>" \
  "https://<elasticsearch-hostname>/"

Use the reported version only for validation and compatibility checks.

Do not use it as a reason to perform direct package or infrastructure changes.

If a workload has an upgrade issue

Identify whether the issue affects:

  • platform availability;
  • Elastic Agent or Fleet;
  • an OpenTelemetry integration;
  • an application client;
  • authentication;
  • indexing or search behavior.

For assistance:

  • AI chat: https://copilot.ivedha.cloud
  • Support portal: https://support.ivedha.com/

Provide the deployment reference, affected workload, and relevant non-sensitive error information.

Do not include passwords, API keys, enrollment tokens, or other secrets.

Elasticsearch downgrade

Elasticsearch does not support a normal downgrade to an earlier version.

If a serious issue occurs after an upgrade, recovery is handled as a managed service recovery process rather than by manually installing an older Elasticsearch version.

Next step

Continue with renew TLS certificates.