We have a web app and use Autoscale running between 1 and 3 instances, scaling on cpu. All works well. What’s incredibly frustrating is that on scale in, the newest server is shut down and we would like the oldest server - now with the highest memory usage, to be closed down. There are no settings in the portal I can see that handle this. Does anyone know if there are controls or additional rules on what server can be shut down first ?
Asked
Active
Viewed 78 times
2
-
Have you considered creating a new autoscale profile and applying it to the scale set? – Citizen Jan 04 '20 at 02:13
1 Answers
0
It is not possible to amend the scaling process for web apps, it will always scale in the newest instance. Web appliacations do not give you much control over the scaling process.
If you wanted more control you could look at using VM Scale Sets instead, but that has the downside of having to manage the VMs. Scale sets provide the ability to have scale profiles.
Sam Cogan
- 38,736
- 6
- 78
- 114