AWS scaling up/down

What is the best way to scale up/down workers on AWS?

In most cases you want to run multiple worker instances one in each container in order to have redundancy. Workers are stateless, so if one container goes down your workflow/activity executions would be able to continue on a different worker.

As far as scaling goes, Temporal itself does not do any auto-scaling for workers. To know when to scale look at the worker tuning guide, and specifically SDK metrics mentioned there (especially workflow_task_schedule_to_start_latency and activity_schedule_to_start_latency) that would give you a good indication on when to scale up workers.

1 Like

Hi Tihomir,

Are these metrics available at cluster level or Namespace? if not at name space. How can I scale, while having multiple namespaces?