Troubleshoot by symptom
Work from the symptom and change one thing at a time. During a live production, prefer the recovery action with the smallest blast radius.
I Can’t Connect to a Machine
Section titled “I Can’t Connect to a Machine”- Confirm the correct organization and environment are active.
- Confirm the machine reports
running. - If it was just started, wait for the DCV session to become ready and retry.
- Try the other supported connection type: Web Interface or NICE DCV.
- Ask an administrator to confirm your Instance Manager and studio permission is active.
My Machine Won’t Start
Section titled “My Machine Won’t Start”- Check whether the machine is already in
pending,starting, or another transitional state. - Wait for the current request to finish before issuing another one.
- Read the complete error shown by Studio. AWS capacity failures may be retried in smaller groups.
- If capacity remains unavailable, ask an administrator whether another approved instance type is suitable.
Do not resize a production machine without confirming application and performance requirements.
An Input Has No Video
Section titled “An Input Has No Video”- Confirm the flow engine reports
ready. - Confirm the flow and input are started.
- Compare the configured protocol and endpoint values with the source.
- Check whether the input status and flow status disagree.
- Verify the source outside the final output path where practical.
Then follow the protocol-specific checks below.
An SRT Connection Won’t Establish
Section titled “An SRT Connection Won’t Establish”For a connection that does not establish:
- Confirm the endpoints use a valid pair: caller and listener, or rendezvous and rendezvous.
- Compare the remote address and UDP port on both ends.
- Confirm firewall, NAT, and cloud security-group rules permit the selected mode.
- Compare passphrase and key length exactly when encryption is enabled.
- Compare Stream ID values when the receiver uses them for access or routing.
Increasing latency does not fix an address, port, firewall, mode, encryption, or Stream ID mismatch.
See SRT settings for connection-mode, latency, and maximum-bandwidth guidance.
NDI Sources Aren’t Visible
Section titled “NDI Sources Aren’t Visible”- Confirm an NDI Discovery Server is deployed, running, and reachable.
- Confirm the instance is configured with the correct Discovery Server address.
- Confirm the NDI source name matches the Flow Manager input.
- Compare NDI send and receive groups where groups are used.
- Confirm the flow engine is ready and supports NDI in the selected direction.
Supported deployment templates can use CLOUDflex Agent to write the host’s NDI discovery configuration.
Video Is Stuttering or Delayed
Section titled “Video Is Stuttering or Delayed”- Check available bitrate and uptime statistics for interruptions.
- Compare configured bitrate or maximum bandwidth with available network capacity.
- For SRT, confirm the latency is appropriate for the actual network path.
- Check whether the issue exists at the source, in the routed output, or only in the viewer.
- Avoid changing several encoding or network values at once.
An Operator Has the Wrong Permissions
Section titled “An Operator Has the Wrong Permissions”- Confirm the operator accepted the organization invitation.
- Confirm they are in the correct organization.
- Confirm an environment permission exists for the required application.
- Confirm the required studio is selected, or that application-wide access was intended.
- Check the permission start and end dates.
Organization membership and environment permissions are separate. A Member role alone does not grant application access.
A Resource Is Still Running
Section titled “A Resource Is Still Running”- Check the Dashboard and Instance Manager in the active environment.
- Confirm whether the resource is intentionally supporting another studio or production.
- Stop it if it is safe to do so.
- Wait for the machine to report
stopped. - Review other environments and Usage if activity remains unexplained.
What to Collect Before Contacting Support
Section titled “What to Collect Before Contacting Support”- Organization and environment names, without credentials or account identifiers.
- Resource or flow name.
- The action attempted and approximate time, including time zone.
- Current state and the complete error message.
- Whether the problem is repeatable.
- Protocol and caller/listener direction where relevant.
- The checks already completed.
Never send passwords, access keys, stream passphrases, private URLs, or tokens in a support screenshot or message.