AWS
Common quotas/limitations
It’s important to understand the quota/limitation of the cloud service before you decide to use it.
e.g. If your API’s response time is > 30s, you should not take API gateway as your solution.
| Service | Resource | Limit |
|---|---|---|
| Lambda | Function timeout | 900 seconds |
| Lambda | Concurrency (regional-level) | 1000 (default) |
| SQS | Message Size | 256 KiB |
| SQS | Max Delay | 15 minutes |
| SQS | Retention | min=60s, max=14d |
| API Gateway | Request timeout | 30 seconds |
| API Gateway | WAF | Only REST API can attach WAF |
| API Gateway | Request per second (rps) | 10000 (per account, maybe various in different region) |
| SSM parameter store | Transactions per second (tps) | 40 (Standard version, per account and region) |
| S3 | Static website hosting | Does not support HTTPS |
| EKS | Max pods per instance (EC2) | The max pod count is various on different EC2 |
EC2 instance spec limitations
Common limitation / tips
- Max network traffic bandwidth
- t-series instances have burst (credit) limitation, must be aware about the credit usage.
- spot-instance / reserved instance / saving plan to save cost.