[steelaz]

This is cool product, but but I don't undersatand this statement:

"Importantly, Akita did not impact processing loss or extra costs inside AWS, a main concern at our company stage."

Author specifically talks about AWS Fargate and links to Akita docs where it says in AWS Fargate setup, Akita agent should run as a side car in each container you deploy. How can that not bring significant amount of extra compute cost?

[ebbp]

Not from the company but we do something similar with opentelemetry. It’s true, because you pay for the total allocation of CPU/memory on Fargate, so you can add a sidecar container into that total allocation with a small deduction from the amount left available to the app itself. E.g. Before: 512MB for task, 512MB available for the application After: 512MB for task, 412MB available for the application, 100MB available for sidecar

[steelaz]

Yes, but even in your example, that's 20% of resources going to the sidecar. Not to mention sizing correctly multiple containers in a single task gets complicated.

[jeanyang]

Hello, this is Jean from Akita. You're right that running Akita as a sidecar does incurs extra compute. Whether or not this incurs extra cost on Fargate depends on whether it's necessary to switch to a larger instance. In zMatch's case, the answer is no. In general, it depends on how much traffic you're sending through Akita.

Something to point out is that Akita passively watches traffic and doesn't sit in the path of traffic, so there's no impact on latency, unless the task is at 100% CPU utilization.