Can anyone clarify BYOC Edge SIP 488 failures at 200 concurrent calls in US1?

Voice & Telephony
1

Can anyone clarify the expected behavior for BYOC Edge capacity limits when using JMeter for load testing?

Background

The environment is Genesys Cloud US1. The team is using JMeter 5.6 to simulate outbound call volume. The goal is to test stability at 200 concurrent SIP sessions. The configuration uses a single BYOC Edge instance connected to our SIP trunk provider. The JMeter script sends SIP INVITE requests to the Edge IP. The script includes a 5-second think time between calls for each thread group. The authentication uses SIP digest credentials. The Edge version is 2024.1.0.

Issue

When the test reaches 150 concurrent calls, the system works fine. The calls connect and audio flows. However, when the load increases to 200 concurrent calls, the SIP INVITE requests start failing. The Edge returns a 488 Not Acceptable Here response. The error happens immediately after the INVITE is sent. The JMeter sampler shows the failure within 200ms. The SIP trunk provider logs show no incoming INVITEs for the failed calls. This suggests the Edge is rejecting the requests before sending them to the trunk. The standard trunk connection works fine at this volume. The issue is specific to the BYOC Edge.

Troubleshooting

  1. Checked the Edge logs. The logs show “SIP invite rejected: insufficient resources” at the time of the 488 errors.
  2. Verified the SIP trunk bandwidth. The trunk supports 500 concurrent calls. There is no limit on the provider side.
  3. Checked the Genesys Cloud admin console. The BYOC Edge status is green. No alerts are present.
  4. Reduced the concurrent calls to 100. The test runs successfully without errors. This confirms the issue is related to the concurrency level.
  5. Checked the API rate limits for the Edge. The GET /api/v2/telephony/providers/edges endpoint shows normal response times. The issue is with the SIP media path, not the control API.

Is there a specific license limit or configuration setting for BYOC Edge concurrent calls? The documentation mentions throughput but does not specify hard limits for small instances. The team needs to know if a larger Edge instance is required for 200 concurrent calls. The current setup seems to hit a wall at 150 calls. Any guidance on capacity planning for BYOC Edges would be helpful.