E2E CloudTIR Auto Scaling Group FAQ
For fixes to a problem you are seeing, see Troubleshooting TIR Auto Scaling Groups.
Basics
Section titled “Basics”What is an Auto Scaling Group?
Section titled “What is an Auto Scaling Group?”An Auto Scaling Group runs a set of identical VMs and grows and shrinks automatically, between Minimum VMs and Maximum VMs. It is a resource of its own, with its own pages. The TIR Auto Scaling Group overview describes the product.
What is a group built from?
Section titled “What is a group built from?”A snapshot of one of your VMs. Every VM in the group is a copy of that snapshot, with the same operating system, packages, and files as when the snapshot was taken. Pick a VM on Select Source, and TIR takes the snapshot for you, or pick a snapshot you already have. See Create an Auto Scaling Group.
Can I build a group from an image?
Section titled “Can I build a group from an image?”No. A group is built from a VM or from a snapshot of a VM.
Does a group use GPU or CPU?
Section titled “Does a group use GPU or CPU?”CPU only. The create form shows CPU plans, and the source must be a CPU VM or a snapshot of one. Every VM in a group uses the one plan you choose at create, and the plan must be billed by the hour.
Which regions can I use?
Section titled “Which regions can I use?”You can use Delhi and Chennai. Auto Scaling Groups is in the side menu under Compute.
Who can use Auto Scaling Groups?
Section titled “Who can use Auto Scaling Groups?”Anyone whose role grants the permission for Virtual Machines and Containers (Instances). Auto Scaling Groups use that same permission. You need create permission to create a group, update permission to change one, and delete permission to delete one. See TIR Access Control.
Scaling
Section titled “Scaling”How many VMs can a group have?
Section titled “How many VMs can a group have?”From 1 to 60. Minimum VMs starts at 2 and Maximum VMs at 5, and the form needs a maximum at least one above the minimum. See Scale on usage.
How does a group decide when to add or remove VMs?
Section titled “How does a group decide when to add or remove VMs?”It follows its policy. A Usage-based group adds VMs when usage stays above your target and removes them when usage stays below it, always within Minimum VMs and Maximum VMs. A Scheduled group changes its capacity only at the times you set. See How auto scaling works.
What can a Usage-based group watch?
Section titled “What can a Usage-based group watch?”CPU utilization, Memory utilization, CPU + Memory utilization, or Network throughput. The group measures the average across all its VMs. See Scale on usage.
Can I use usage and a schedule together?
Section titled “Can I use usage and a schedule together?”Yes. Add schedules to a Usage-based group, and each schedule sets the minimum the group works above for its window. To combine them, keep the Usage-based policy and add schedules to it. See Scale on a schedule.
What time zone do schedules use?
Section titled “What time zone do schedules use?”UTC. The Runs at time is a cron expression in UTC, so convert your local time to UTC when you enter it. TIR spells the time out in plain words, with the next run. See Scale on a schedule.
How long does a group wait before it adds or removes a VM?
Section titled “How long does a group wait before it adds or removes a VM?”Usage has to stay past the target for the whole stabilization time. The Balanced preset, the default, waits 150 seconds before adding and 600 seconds before removing. Responsive waits 60 and 300 seconds, and Cost optimized waits 300 and 1,200 seconds.
What is Scale Now?
Section titled “What is Scale Now?”Scale Now asks a Running group to run a number of VMs right away. On a Usage-based group, a count above the saved minimum is temporary: the group holds it for a short time, then goes back to its own range. To keep a higher number, raise Minimum VMs from Edit. See Use Scale Now.
Can I change the plan of a group?
Section titled “Can I change the plan of a group?”No. The plan, the disk size, and the source snapshot are fixed when you create the group, because changing them would mean replacing every VM. You can change the name, the capacity, the policy, the schedules, and the health checks. To get a different plan, create a new group, for example with Clone. See Edit an Auto Scaling Group.
Health and replacement
Section titled “Health and replacement”What happens to a VM that is unhealthy?
Section titled “What happens to a VM that is unhealthy?”The group always watches for VM and host failures, and it replaces a VM that was running and then fails. You can also add an HTTP or TCP check of your own application. The group builds the new VM from the source snapshot. See Set up health checks.
Data and access
Section titled “Data and access”What happens to the data on a VM when the group removes it?
Section titled “What happens to the data on a VM when the group removes it?”Treat it as lost. A new VM starts from the source snapshot, so data written on a VM after it started exists only on that VM. To keep data, store it outside the VM. See How auto scaling works.
Do changes I make on the source VM reach the group?
Section titled “Do changes I make on the source VM reach the group?”No. Every VM is a copy of the snapshot as it was when taken. To use later changes, take a new snapshot and create a new group.
What if I delete the source VM?
Section titled “What if I delete the source VM?”The group keeps working. It holds its own copy of the plan, disk size, SSH keys, and source name, so Retry Launch still works too. Clone is available only while the snapshot itself exists.
Does every VM get a public IP address?
Section titled “Does every VM get a public IP address?”Yes. Every VM gets its own public IP address once it is ready, and the address is released when the VM is removed. A load balancer gives the group one shared address. See Manage SSH keys and Security Groups.
How do I sign in to a VM of the group?
Section titled “How do I sign in to a VM of the group?”Use SSH as the root user with the VM’s public IP address: ssh root@<public IP>. The command is under
Connect on Overview, and the group needs an SSH key and a Security Group that opens port 22. See
Manage SSH keys and Security Groups.
Can I use a load balancer?
Section titled “Can I use a load balancer?”Yes, a Web traffic (HTTP/HTTPS) load balancer. See Connect a load balancer.
Billing
Section titled “Billing”What does a group cost?
Section titled “What does a group cost?”Each VM is billed at the hourly price of the plan you chose, plus its disk above the first 30 GB. The cost follows the number of VMs the group runs. For prices, see E2E Networks pricing. The create and edit forms show an Estimated cost before you save. See Billing for TIR Auto Scaling Groups.
Is a VM billed while it starts?
Section titled “Is a VM billed while it starts?”No. Compute is billed by the minute, and only from the moment the VM is ready until it is removed.
Am I billed while a group is stopped?
Section titled “Am I billed while a group is stopped?”Compute billing for its VMs ends when you stop the group. Each VM’s disk above the free 30 GB keeps billing while the group is stopped. See Stop and start an Auto Scaling Group.
Who receives the bill?
Section titled “Who receives the bill?”The owner of the project’s team, even when another member created the group.
Does deleting a group delete the source snapshot?
Section titled “Does deleting a group delete the source snapshot?”No. Deleting a group terminates the group’s VMs. The source snapshot, the source VM, the SSH keys, and the Security Groups stay. See Delete an Auto Scaling Group.
Monitoring
Section titled “Monitoring”How do I see what the group did?
Section titled “How do I see what the group did?”Open the group’s Scaling Activity tab. It lists every scaling change and configuration change, newest first, with the reason and who or what did it, and you can export the list as CSV or JSON. See View scaling activity and logs.