Rate Limits & Throttling

Aevum Zenth's API enforces rate limits to ensure stability, security, and fair resource allocation across all 400+ subsidiary endpoints. This guide outlines global limits, header specifications, and best practices for production integrations.

Global Rate Limit Tiers

Rate limits are applied on a sliding window basis. Exceeding limits returns 429 Too Many Requests. Enterprise clients receive dedicated infrastructure and custom throttling policies.

Rate Limit Response Headers

Every API response includes standardized headers to help your application track usage and implement adaptive throttling. These headers are consistent across all Aevum Zenth endpoints.

X-RateLimit-Limit: 600
X-RateLimit-Remaining: 482
X-RateLimit-Reset: 1700982400
Retry-After: 12 (only present on 429 responses)
X-AZ-Request-Id: az-req-8f3d9c2a-4b1e-4f8a-9d2c-7e6f5a4b3c2d
X-AZ-Quota-Pool: shared-gateway-v4

Header Specifications

• X-RateLimit-Limit: Maximum requests allowed in the current window.
• X-RateLimit-Remaining: Requests left before hitting the limit.
• X-RateLimit-Reset: UNIX timestamp when the quota window resets.
• Retry-After: Seconds to wait before retrying (only on 429).
• X-AZ-Request-Id: Unique trace ID for debugging and support tickets.

Endpoint-Specific Overrides

Certain high-compute endpoints enforce stricter limits to protect downstream services. When an endpoint-specific limit is lower than your global tier, the stricter limit applies.

Recommended Retry Strategy

To maximize throughput and avoid cascade failures, implement an exponential backoff with jitter. The Aevum Zenth SDKs handle this automatically, but manual implementations should follow these guidelines:

if response.status == 429:
    wait_time = response.headers["Retry-After"]
    jitter = random.uniform(0, 0.5 * wait_time)
    sleep(wait_time + jitter)
    if retry_count < 3:
        retry_count++
        return make_request()
    else:
        throw RateLimitExhaustedError()
elif response.status >= 500:
    backoff = min(30, 2 ** retry_count) * 1000
    sleep(backoff + random_jitter())
    retry_count++

Monitoring & Quota Management

Track your consumption in real-time through the Aevum Zenth Developer Console. Set up webhook alerts for 90% quota utilization to prevent unexpected throttling.

• Dashboard Metrics: View requests/min, error rates, and latency percentiles across all endpoints.
• Quota Reserves: Pro/Enterprise plans can purchase burst credits for peak-load periods.
• Support Triage: Include X-AZ-Request-Id when reporting rate limit issues. Our SRE team can trace gateway nodes within 60 seconds.