Octopus Cloud unable to communicate to deployment target vs a proxy

[kieran.hayes]

My company has an Octopus Server on-premise that has been in use for many years. Im now tasked with helping to move this to Octopus Cloud.

With the Octopus Server on-premise it has a use-case to deploy into an Azure VM that is private. It does this via proxy. That works fine.

I now have a brand new Octopus Cloud instance but I am unable to get it to talk to the same deployment target via the same proxy. The connectivity always times out after 1 minute.

I have confirmed that the proxy is publicly available over the internet and I can issue a curl request to the deployment targets tentacle via the proxy and that works fine.

I have also updated the trust on the deployment targets tentacle so it trusts both the old Octopus Server as well as the new Octopus Cloud instance. This was done via the following command and restarting the tentacle.

.\Tentacle.exe configure --instance “Listening Tentacle” --trust MYTHUMBPRINT

It is unclear to me why the Octopus Cloud instance can not communicate to the proxy, and I have no means to troubleshoot further. My assumption here is that the in-built worker has no outbound access to the Internet, but I’ve no means to verify that.

Using a custom worker doesn’t really solve any problems here because we wouldn’t be able to use a worker that is accessible over the Internet. Meaning we would need to use the same proxy that Octopus Cloud is already failing to connect to.

How is it that an Octopus Cloud instance is meant to connect to a proxy? How can I troubleshoot further why it can not connect?

[kieran.hayes]

Oddly as soon as I posted this the connectivity does work… Not sure if there was transient Azure network outage or what.

[clare.martin]

Hey @kieran.hayes,

Thank you for letting us know this was fixed, Azure did have an outage earlier which was intermittent as it seemed like someone cut a fibre cable so there was brief periods where one region was moved to another to prevent a full outage.

We had an email from Azure this morning:

At 22:05 UTC on 14 Sep 2023, we experienced a fiber cut that impacted traffic traversing between Availability Zones (AZs) in the West Europe region. During this time, customer applications may have experienced periods of 1 to 5 minutes of latency or timeouts for traffic that was traversing between AZs in the region. Traffic moving in or out of the region or traversing within AZs was unaffected.

Not sure if your Azure VM would have been affected by this but it may explain the sudden connection success having not changed any settings, just wanted to let you know in case you were scratching your head as to why this suddenly just started working.

If that outage would not have affected you please ignore this email,

Kind Regards,
Clare