I've got two database instances, a db.r7g.2xlarge which is a master Aurora MySQL, and a db.r7g.large which is a read-only instance used for some processing. I purchased reserved instances for both the large and 2xlarge, but when I check my billing, it seems like it's not applying the reserved instances properly. Instead, it uses the large reservation for the 2xlarge, which ends up charging me an extra $320 a month, while it only partially uses the 2xlarge reservation for the large instance. I'm really confused and frustrated, as it feels like a bug in the system. AWS support says this is how it's supposed to work, but it doesn't seem right since I'm actually using both instance types. Anyone else dealing with the same issue?
2 Answers
Are your instances configured as Multi-AZ? That could be a factor in how your reserved instances (RIs) are utilized. Just a thought!
Make sure your reserved instances are set to Regional Scope with Size Flexibility if that option is available. This way, your RIs can be applied across the region to any compatible instance. If your workloads are fixed, consider purchasing specific RIs for each instance size you’re using.
Exactly what I've done! Both my instances should match the reserved ones, but they’re still not being applied correctly.
Actually, Aurora doesn’t support Multi-AZ configurations like that. The writer and reader instances are in different AZs but are still single instance types.