Replies: 1 comment 1 reply
|
Your rollout plan is broadly the right shape, with two important additions. First, diff the complete release manifest, not only the operator Deployment. An operator release can change RBAC, ServiceAccount, webhook, or related resources even when the Jaeger CRDs are byte-for-byte identical. Save both the rendered old and new manifests, then run kubectl diff before applying. Second, verify compatibility for the Kubernetes version and the Jaeger generation you are running. The operator repository keeps a compatibility matrix here: https://github.com/jaegertracing/jaeger-operator/blob/main/COMPATIBILITY.md A conservative production sequence is: Then verify reconciliation, collector/agent ingestion, query, and storage separately. Keep the old manifest so reverting the operator image is quick; do not delete CRDs as a rollback step. One caveat: the current operator README marks this operator as deprecated and directs Jaeger v2 users to the OpenTelemetry Operator. That does not prevent a controlled Jaeger v1 operator upgrade, but it is worth confirming that the production workload is still on the operator-supported generation: https://github.com/jaegertracing/jaeger-operator/blob/main/README.md |
Uh oh!
There was an error while loading. Please reload this page.
Hello,
I am planning to upgrade Jaeger Operator in production from
1.62.0to1.65.0.I compared the official manifests. The only difference is:
All reactions