Skip to main content

Quality of Service

Quality of Service (QoS) in Connect protects your APIs from cascading failures and long-running downstream calls.

It improves system stability by:

  • Preventing repeated calls to failing services
  • Failing fast when downstream systems are unresponsive
  • Protecting upstream clients from excessive wait times

QoS ensures your platform remains responsive even when backend services degrade.

Quality of Service configuration screen showing circuit breaker and timeout settings

Figure 1: Quality of Service configuration interface in Connect.

Where to Configure​

Navigate to:

Service → Settings → Quality of Service

Click the Enabled toggle to activate QoS enforcement for the selected service.

How It Works​

Connect applies QoS policies in two ways:

  1. Timeout Control -- Limits how long a request can wait for a downstream response.
  2. Circuit Breaker -- Temporarily stops calls to a failing service after consecutive errors.

These mechanisms prevent cascading failures and reduce system strain.

Circuit Breaker​

The Circuit Breaker protects downstream services from repeated calls during failure conditions.

Configuration​

FieldDescription
Exceptions ThresholdNumber of consecutive failures required to open the circuit
Break DurationTime (in seconds) the circuit remains open before attempting recovery

Example​

Exceptions Threshold: 3
Break Duration: 10 seconds

This means:

  • After 3 consecutive failures, the circuit opens.
  • All further requests fail immediately without calling the downstream service.
  • After 10 seconds, Connect allows a trial request.
  • If the trial succeeds, the circuit closes.
  • If it fails, the circuit opens again.

Behavior​

When the circuit is open:

  • The downstream service is not called.
  • Requests fail immediately.
  • This protects the system from cascading failures.

The circuit breaker is based on consecutive failures, not total failures over time.

Timeout Control​

Timeout Control limits how long Connect waits for a downstream response.

Configuration Field​

FieldDescription
Timeout IntervalMaximum time (secs) allowed for a downstream request before it is aborted

Example​

Timeout Interval: 30 seconds

This means:

  • If the downstream service does not respond within 30 seconds,
  • The request is aborted,
  • A timeout error is returned to the client.

Timeouts prevent threads from being held indefinitely and improve overall responsiveness.

Error Responses​

Both Timeout and Circuit Breaker failures return HTTP 503 Service Unavailable.

info

The response body message can be customized in Global Settings.

Enforcement Behavior​

When Timeout Occurs​

  • The downstream call is aborted.
  • An error response is returned to the client.
  • The failure counts toward the circuit breaker threshold.

When Circuit Is Open​

  • Downstream service is not invoked.
  • The request fails immediately.
  • Connect attempts recovery after the configured Break Duration.

QoS enforcement happens at the Gateway Layer, before forwarding requests to downstream services.

Best Practices​

  • Use conservative timeout values based on realistic service latency.
  • Configure a low Exceptions Threshold for critical services.
  • Avoid overly long Break Durations.
  • Monitor logs to detect recurring failure patterns.
  • Combine QoS with Rate Limiting for complete traffic control.