I was skeptical, but after using Q Developer for a month on a Go-based microservices migration, I'm impressed. The Go support is much stronger than I expected, especially for AWS SDK v2 and Lambda.
Key takeaways so far:
* It nails common boilerplate (DynamoDB scan ops, S3 event handling).
* Explains context-specific AWS errors well.
* Surprisingly good at suggesting cost-saving changes (e.g., "use Lambda power tuning for this batch job").
Biggest win? It correctly refactored a messy `http.Handler` into a cleaner structure with proper error wrapping. The suggestions were actually usable.
My main question for others: **What's the actual ROI?** For a team already using Go + AWS, is the time saved worth the per-user cost? I'm leaning yes, but curious about larger team experiences.
—CR
Ask me about hidden egress costs.
Totally agree on the boilerplate and error handling! I experienced the same with some SQS -> Lambda -> DynamoDB flows. It cut my setup time in half for those patterns.
On your ROI question for a larger team, we trialed it with three backend devs for a quarter. The real value wasn't just the time saved on initial code, but the consistency it forced on the codebase. Every lambda handler ended up with the same error logging structure without us having to nitpick in PRs. That alone was worth the cost for us.
The only hiccup we hit was around some VPC configuration hints it gave for Lambda. They were technically fine, but didn't match our specific security group patterns. Just a reminder to always check its infra suggestions against your own setup. 😊
measure twice, ship once
Your point about it explaining AWS errors well is spot on. That's been huge for our junior devs. Instead of just getting "AccessDenied," it often suggests checking the specific IAM action or resource ARN format that failed. Saves a lot of back and forth.
On ROI for a Go/AWS team, our math came out positive. The per-user cost was less than half a day of senior dev time per month. If it saves each person a few hours of debugging or boilerplate, it pays for itself. The consistency bonus user1412 mentioned is real, too.
Have you tried it with any of the Go AWS CDK? I found its suggestions for structuring Pulumi a bit hit or miss.
> "it often suggests checking the specific IAM action or resource ARN format"
Yes! This is so helpful when you're learning. I've wasted hours on vague "AccessDenied" errors before. 😅
I haven't tried it with CDK yet, but your comment about it being hit or miss with Pulumi makes me hesitant. Maybe it's strongest on core SDK code and weaker on the infra-as-code layer?