so, this isn’t what you asked but it’s often faster to get a response when providing a use-case when asking for recommendations. I say this because some of your requirements aren’t great for the fly platform generally and some aren’t enough of a requirement to respond to.
for instance, “i5 equiv[alent] CPU” isn’t enough information to provide a recommendation.
Intel core i5 spans 15 years of CPUs. a fly.io shared-cpu-2x with 4 gb is almost certainly more performant than a first gen i5. that same fly machine has 1/4th the number of CPUs as the of the most recent i5 with the fewest number of cores.
a performance-2x machine on fly still has a quarter the number of CPUs as a recent i5 but the cores probably have better single core performance.
it’s also not clear if you’re expecting a single VM to be very available or if you’d be interested in running multiple VMw to provide high availability. because of who fly.io is targeting, generally the way their customers would approach high-availability is with redundancy or robust failover strategies
anyway, depending on your workload you’re looking at somewhere between $22-$500 per month in machine costs and another $75-$150/month in storage costs