Quotas

The Quotas tab displays throughput limits applied to producers and consumers in your cluster, helping ensure fair resource allocation.

Overview

Shows two levels of quota configuration:

• Default quotas: Cluster-wide limits that apply to all clients unless overridden.
• User overrides: Per-principal limits that take precedence over the defaults for specific clients.

Quota Types

Quota	Description
Producer byte rate	Maximum bytes per second a producer can send to the cluster.
Consumer byte rate	Maximum bytes per second a consumer can fetch from the cluster.
Request percentage	Maximum fraction of I/O and network threads a client can use. Limits CPU impact from a single client.

User Overrides

Overrides apply custom limits to a specific principal, identified by username.

They are useful when:

• A high-throughput service needs a higher limit than the cluster default.
• A low-priority consumer needs to be throttled to protect production workloads.
• A Zilla Gateway service account needs elevated throughput for aggregated API traffic.

Tips

If your API Products serve high-volume consumers, consider setting a user override for the Gateway's service principal to avoid hitting cluster-wide default limits during peak traffic.

Interpreting Throttling

When a client exceeds its quota, Kafka throttles it by introducing artificial delays in responses. You may see:

• Increased produce or consume latency for affected clients.
• THROTTLING_QUOTA_EXCEEDED errors in client logs.

Use this view to identify clients approaching or exceeding their quotas and adjust overrides accordingly.