Option to Deploy Across Multiple Regions / Data centres

[taylora3]

I was on the call during the week. I would like a way to deploy across multiple regions. e.g have 2 Moodle deployments in an active/active set up e.g one in North Europe and one in West Europe. The aim is to make sure that the set up is always on. So if the Data centre in North Europe goes down it will fall back to the one in West Europe and vice-versa. What is the projected uptime for the Maximal deployment? Many thanks and happy to work with you.

[hosungsmsft]

Thanks much for your input also attending the webinar earlier this week. We'll definitely consider adding the geo failover support, but I have to tell you that it may not come soon. We definitely welcome any contribution from the community, and we'd be very happy to work with you on this feature as well, so please feel free to send PRs or leaving more comments/issues here.

Regarding your question, I'm not sure what you mean by the "projected uptime for the maximal deployment". Did you mean the SLA (like xx.x %) of a maximal deployment?

[taylora3]

@hosungsmsft You are right. I meant the SLA. e.g the 99.99% up-time referenced in azure availability zones We are looking at worst case scenarios and disaster recovery. If you could point me in the right direction of docs etc that would be appreciated.

[hosungsmsft]

Thanks for your clarification. Due to the "solution"-nature of our templates, we currently don't have a concrete SLA number of a deployed Moodle cluster as a whole. We may try to do some reliability analysis of an entire cluster from individual components' SLA numbers, but that may not be very accurate or useful. Component-wise SLAs are as follows for the NFS option (no file system high availability):

• Controller VM: 99.9%, as it's a single instance VM and according to this Azure VM SLA documentation.
• VMSS VMs: 99.95%, as they are deployed in an Azure availability set. If they are deployed in availability zones (which are said to be in preview on select regions only), SLA will be 99.99% (see the same Azure VM SLA documentation above), but I don't think our templates currently support the option of deploying VMSS VMs in availability zones. We'll look into that.
• MySQL DB: 99.99% according to this Azure MySQL SLA documentation.

If customer chooses Gluster for file system high availability, an Azure availability set-based Gluster VM cluster is also deployed, whose SLA is 99.95%. The maximal deployment also deploys an Azure Redis cache, whose published SLA is 99.9%. Please note that there are other components' SLAs that'll affect the SLA of the entire cluster as a whole. Hope this helps.

[taylora3]

Thanks for your help. I'll see if that satisfies my clients needs. I'm looking in to global vnets as a possible solution if it has to span across regions to achieve always on.