Chasing quick wins on cloud spend

Quick question: if you had to show savings in the first 60 days, would you prioritize rightsizing compute with AWS Compute Optimizer or commit to a 1‑year Savings Plan at about 30% discount? I’ve modeled both for a client with a steady 120 vCPU baseline and the break-even shifts fast if workloads spike on Fridays; curious what’s worked for you without locking in waste.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍​⁠‌‍⁠​‌‍⁠⁠‌⁠‌‌‌‍‌​‌‍​⁠‌‍⁠⁠‌‍⁠‌‌⁠​​‌⁠‌‌‌⁠‌​‌‍‍‌‌‍⁠‍‌‍‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‌​⁠​​​⁠‌‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍​⁠‌​​⁠‌‌​⁠‌‌‌‌​⁠‌⁠‌‌‍‌‌⁠‍‌‌​‌‌‌‌​⁠‌​‌‍‌‍⁠​‌‌‍‌‌​​‍‌‌‌​‌‍‍​‌​‌​​‍​‍‌⁠⁠‌

I’d cap commitment to about 60–70% of the “120 vCPU baseline” on a 1‑year no‑upfront Compute Savings Plan after a quick Compute Optimizer pass, which has given me about 20% in month one while keeping Friday spikes on on‑demand. If those spikes last more than a few hours, I’d lower coverage further and lean on Spot via your ASG — how long do Fridays run?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍​⁠‌‍⁠​‌‍⁠⁠‌⁠‌‌‌‍‌​‌‍​⁠‌‍⁠⁠‌‍⁠‌‌⁠​​‌⁠‌‌‌⁠‌​‌‍‍‌‌‍⁠‍‌‍‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‍​⁠‌‌​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‌​⁠​​​⁠‍​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌⁠​​‌​‍‍‌‌​‍‌​⁠‌​⁠‌‍‌​‍‍‌​⁠​‌‍​⁠‌​​‍‌​‌‍‌⁠‌⁠‌​‍⁠‌​‌⁠‌‌‍‌‌‍‌‌​⁠‌​​‍​‍‌⁠⁠‌