Environment:
- Headwind MDM self-hosted on a dedicated VM (hosting provider plan includes 100TB/month traffic allowance)
- Enrolled devices: ~1,500–1,800 Android devices (sizing was originally planned for 5,000+ devices)
- Server has been taken offline by our hosting provider because we exceeded the 100TB monthly traffic quota
Issue:
Our monthly traffic consumption is far higher than expected for our current device count. Based on the provider's traffic graphs (the only data we currently have access to, since the server itself is offline), the vast majority of traffic is generated in the device → server direction, split between:
- Port 443 (HTTPS)
- Port 31000 (MQTT)
Since all traffic is encrypted, we are unable to inspect payloads to determine exactly what is generating this volume. In addition, the VM cannot be brought back online until the next billing cycle, so we cannot currently access server-side configuration, logs, or the admin panel to review our current settings (push intervals, policies, location tracking frequency, app catalogs, log/report collection, etc.).
We are also working with our hosting provider in parallel to try to restore access to the VM sooner, but for now we have no way to check our own configuration.
Questions for the Headwind team:
- What are the most common causes/misconfigurations that lead to abnormally high MQTT (31000) and/or HTTPS (443) traffic between Android devices and the server?
- Which specific settings/policies should we check first once we regain access (e.g. location update interval, heartbeat/keep-alive frequency, log upload frequency, screenshot/screen recording features, app catalog sync, file transfer, remote control sessions, etc.) that are known to drive high bandwidth usage per device?
- Do you have any internal benchmarks or test results regarding expected traffic per device/per month under normal usage, so we can compare against our actual consumption once monitoring is possible?
- Are there recommended best practices or configuration guidelines specifically aimed at minimizing bandwidth consumption for large-scale deployments?
Any guidance you can provide — even before we regain full server access — would be extremely helpful, since we need to identify and fix the root cause before the VM is restored to avoid immediately exceeding the quota again.
Thank you in advance for your support.