Requirements:

• HTTP Version: HTTP messages must include the protocol version in the request or status line (e.g., 'HTTP/1.1').
• HTTP/1.0: If using HTTP version 1.0, only use Content-Length to specify the message body size.
• HTTP/1.1 and later versions: Prefer Transfer-Encoding: chunked for more efficient data transmissions but ensure that Content-Length is not present simultaneously.

• Definition: The Content-Length header indicates the size of the message body in bytes.
• Usage:
  • Mandatory for response messages containing a body, unless Transfer-Encoding is used with the value chunked.
  • Must not be used if Transfer-Encoding: chunked is present.
• Accuracy:
  • Clients must ensure that the size specified in Content-Length matches the size of the received message body.

• Definition: The Transfer-Encoding header specifies the encoding(s) used to transfer the message body. Each encoding is applied sequentially, in order, to decode the message body.
• Values and Usage:
  • chunked: Allows for the transfer of data in variable-sized segments, which is useful when the total length of the message body is not known beforehand.
  • If Transfer-Encoding: chunked is used, the Content-Length header must be omitted.
  • The value chunked must be the last encoding if multiple encodings are used.

• Target Availability: 99.16% monthly
• TDI Monitoring: GET /GetSystemStatus every 1min, timeout 3s
• States:
  • ✅ Healthy: all critical components OK.
  • ⚠️ Degraded: service limited but functional.
  • ❌ Unhealthy: critical components down or timeout.
• SLA Calculation: only unhealthy states count as downtime

• GenerateLabel: 500 ms – 1000 ms
  • Measured internally within our system. Excludes client-side network latency, request transmission, and response delivery times.
  • Carrier API performance can significantly impact this metric.

• GenerateLabel:
  If shipment processing thresholds (number of call per hour) are exceeded, an internal alert is automatically triggered within our system.
  • 00:00 – 03:59: 200 calls/minute per sender
  • 04:00 – 23:59: 100 calls/minute per sender
  (CET/CEST timezones)

⚠️ Important: Avoid parallel operations on the same masterId whenever possible.

• Same method (multiple GenerateLabel calls):
  • Technically supported via a queue mechanism.
  • ⚠️ Requests are processed sequentially under the same masterId, which may cause performance degradation when multiple calls are made in parallel.
  • Recommendation: Prefer sequential calls for the same masterId to ensure optimal performance and predictable processing times.
• Different methods (GenerateLabel + DeleteLabel):
  • ❌ Strongly discouraged and not supported.
  • No concurrency handling implemented.
  • May result in errors and/or inconsistent states.

• Any load testing must be conducted with mandatory support from our technical team.