Update Runtime Parameters Without Redeploying a Job

You can safely and quickly update selected runtime parameters of a running deployment without stopping or redeploying the job. The Dynamic Parameter Updates feature lets you adjust operational settings such as parallelism and checkpoint configurations while your job continues running, allowing updates to take effect within seconds.

Traditional configuration changes require stopping and restarting the deployment, which interrupts traffic, triggers state backfills, or causes short-term resource spikes.

With dynamic parameter updates, VERA applies supported changes in place. The JobManager and TaskManagers stay in use, and depending on the change, VERA either performs a lightweight operator restart or applies the update without restarting. This approach reduces downtime and enables quick, safe scaling of TaskManagers as well as tuning of checkpoint configuration.

How It Works

With dynamic parameter updates, VERA lets you apply changes in place without redeploying the entire job. This approach reduces downtime and allows jobs to continue running with minimal disruption.

This process has three steps:

1. Make changes: Update one or more supported parameters in the deployment configuration.
2. Validate changes: VERA checks which updates can be applied dynamically.
3. Implement changes: VERA applies the eligible updates in place without replacing the JobManager or TaskManagers:
   • Performs a lightweight operator restart if required.
   • Applies the update without restarting when safe.
   • Uses resource pre-application and lazy state loading to minimize interruptions.

The result is a shorter pause for data processing compared to a full redeploy.

Supported Parameters

You can dynamically update only the parameters listed below. Changes to any other parameter require a standard restart to take effect.

Parameter	Description
Parallelism	Controls the number of parallel instances for TaskManagers and operators, enabling dynamic scaling.
Checkpointing interval	Sets the target interval between consecutive checkpoints.
Checkpoint timeout	Defines the maximum allowed time for a checkpoint to complete.
Minimum pause between checkpoints	Specifies the minimum time gap between successive checkpoints.

Important Usage Notes

This feature is supported on VERA 4.1 or later and is experimental. Brief service interruptions may occur while changes are applied.

Keep the following points in mind when applying dynamic updates to ensure safe and predictable results.

• Dynamic updates to parallelism are not supported in Expert mode (fine‑grained resource mode).
• Dynamic scaling applies only to operators whose parallelism is not explicitly fixed in code or configuration:
  • Examples of operators that require specific parallelism include a global operator that must be 1, or a Kafka source limited by partition count.
  • Examples of operators with hard-coded parallelism include those using DataStream#setParallelism(...) or fixed source/sink settings.
• Dynamic updates are evaluated per change. If a change set mixes supported and unsupported parameters, only the supported ones are applied dynamically, and unsupported parameters require a standard restart.
• Sources and sinks must be able to handle the new parallelism (for example, Kafka partitions ≥ desired source parallelism).
• Typical interruptions last ~5 seconds to 1 minute, depending on job graph and state size.

Apply and Manage Updates

You can perform, manage, and, if necessary, roll back dynamic parameter updates.

Apply a Dynamic Update

You can perform a dynamic parameter update using the Ververica Unified Streaming Data Platform Console UI.

1. From the Console, navigate to Operation Center → Deployments.
2. Open the Running deployment you want to update.
3. On the Configuration tab, choose Resources or Parameters, then click Edit.
4. Modify any of the supported parameters: parallelism, checkpointing interval, checkpoint timeout, minimum pause between checkpoints.
5. Click Save.
6. In the upper‑right corner, click Dynamic update (may also appear as Hot update). This applies the changes without a restart.
7. Click Confirm.
8. Watch the deployment status and metrics. A dynamic‑update indicator appears while the change is applied.

Apply Changes to Unsupported Parameters

If you updated any unsupported parameters (see Supported Parameters), you must perform a standard restart for those changes to take effect.

Rollback a Dynamic Update

If a dynamic update negatively affects performance or stability, select one of the steps below to revert safely:

• Revert the parameter to its previous value and perform the dynamic update again.
• Perform a standard restart with the prior configuration to return to a known good state.

Troubleshooting

If the Dynamic update button is not visible:

1. Confirm you modified only supported parameters. (Review Supported Parameters.)
2. Ensure the deployment is Running and your engine version is VERA 4.1 or later.

If the update takes longer than expected:

Large keyed state or complex topologies increase pause time. Consider one or more of the following options:

• Reduce state size.
• Enable or increase incremental checkpoints.
• Apply changes during low‑traffic windows.

If a parallelism change is not applied to all operators:

1. Check for operators with explicit parallelism in code or connector configuration (for example, a fixed source/sink parallelism or a global operator).
2. Remove or relax those constraints if dynamic scaling is desired and safe.