[ECS] [request]: Expose service auto-scaling min/max configuration as Cloudwatch metrics
Community Note
- Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
- Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
- If you are interested in working on this issue or have submitted a pull request, please leave a comment
Tell us about your request Send ECS service auto-scaling minimum and maximum configuration values to Cloudwatch
Which service(s) is this request for? ECS
Tell us about the problem you're trying to solve. What are you trying to do, and why is it hard? Due to the spikey nature of traffic that we serve, auto-scaling cannot be relied upon to react quickly enough to scale our systems and keep things operational. Thankfully the spikes are predictable, so we are able to scale up the affected services via Slackbot (talking to AWS API) by setting the "minimum" and "desired" level for the service to a higher number. This works great, but sometimes we have mutiple spikes over the course of a few hours, so we leave it "scaled up" (higher minimum) and then people forget to "scale it down" (ie. set minimum back to normal level) again once all the spikes have passed.
For EC2-based applications, we are able to query the Cloudwatch metric for ASG minimum size (enabled via the "Enable auto-scaling group metrics collection" toggle) and alert if this number has been set higher than a threshold for more than a configured time period. This reminds people to scale back down.
I have just tried to implement the same monitoring of ECS services, but unfortunately it does not appear that minimum (or maximum) service scale is published to Cloudwatch as a metric. It would be great to (potentially optionally like EC2 ASGs) have this information fed to Cloudwatch for observability purposes.
Are you currently working around this issue? Barely. using Slack scheduled messages and eyeballs looking at dashboards.
Additional context Anything else we should know?
Attachments If you think you might have additional information that you'd like to include via an attachment, please do - we'll take a look. (Remember to remove any personally-identifiable information.)