Evaluate ECS cluster resources and create a plan to monitor over time - Githubissues

cityofaustin / atd-data-tech

Austin Transportation Data & Technology Services

18 stars 2 forks source link

Evaluate ECS cluster resources and create a plan to monitor over time #18168

Open mddilley opened 4 months ago

mddilley commented 4 months ago

When working on the Moped 2.14 launch, Frank noticed that we are under-utilizing the resources allocated to the Moped ECS clusters. Ideally, we would have a plan to check on these resources and adjust them as needed to keep an eye on costs, etc. This also has the benefit of giving us a pattern to manage resources for all ECS deployments across all of our products.

Possible tasks:

manual check-ins on resource allocation with a schedule (quarterly?)
- calendar invites? slack alert?
set up AWS alerts to help us know when we need to increase
fix up the health check to unblock autoscaling

There could also be some crossover with https://github.com/cityofaustin/atd-data-tech/issues/11751 which is about health monitoring.

mddilley commented 3 months ago

@johnclary I added the VZ label and updated the issue with an item about the health check & autoscaling.