Title: [Question] Safe way to update snapshot interval without downtime in K8s Operator environment

[kcj-3054]

Describe the bug We are currently operating DragonflyDB in a Kubernetes environment using the Dragonfly Operator. Our current configuration creates snapshots every 5 minutes, which we believe is too frequent and risky for our production workload. We would like to increase this interval (e.g., to 1 hour) or change the schedule.

My primary concern is whether updating the snapshot arguments in the Dragonfly CRD will trigger a Pod restart. If it triggers a restart (Rolling Update), we want to confirm if this process guarantees zero downtime and data consistency during the update. Is there a way to update this configuration dynamically without restarting the pods?

To Reproduce Steps to reproduce the behavior:

Deploy Dragonfly using the K8s Operator with a specific snapshot schedule (e.g., Args: ["--snapshot_cron=*/5 * * * *"]).

Observe that snapshots are taken every 5 minutes.

Edit the Dragonfly CRD to change the snapshot schedule (e.g., change to */60 * * * *).

Question: Does this action trigger a rolling restart of the StatefulSet? If so, is it safe for client connections?

Expected behavior I expect to be able to update the snapshot configuration (interval) without interrupting the service (Zero Downtime). Ideally, I would like to know the recommended procedure to apply this change safely in a production environment via the Operator.

Screenshots N/A

Environment (please complete the following information):

OS: [ubuntu 20.04]

Kernel: uname -a (Please fill this if known, otherwise leave as is)

Containerized?: Kubernetes (managed by Dragonfly Operator)

Dragonfly Version: [Please enter your version, e.g., v1.14.0]

`apiVersion: dragonflydb.io/v1alpha1
kind: Dragonfly
metadata:
name: dragonfly-sample
spec:
replicas: 3

Current configuration with frequent snapshots

args:
- "--snapshot_cron=*/5 * * * *"
# We want to change this to a less frequent interval without downtime.

[kcj-3054]

Hi AiGentsy,

Thanks for your interest in contributing! Could you please share more details about your analysis and the specific solution you have in mind?

We'd like to understand the technical approach before you proceed with a PR.

Best regards,

[romange]

it's a bot. blocked

[romange]

@Abhra303 @miledxz can you please answer?