8 ms·
You're making a tradeoff. You no longer have to alarm for things like your server failing due to disk failure or other stuff that lambda abstracts away from you
by qvrjuec 5y ago
You're making a tradeoff. You no longer have to alarm for things like your server failing due to disk failure or other stuff that lambda abstracts away from you, but you do have to alarm on things that will increase cost. Constant cost(if you aren't dynamically scaling) and more headache vs variable cost and less headache
- paulgb 5y agoIt would be a trade-off I'd consider if alternatives didn't exist, but they do. Someone down-thread mentioned Cloudflare Workers, which have a sensible pricing model. These days I use Google Cloud Run, which gives me everything I wanted from Lambda but has model where you pay by CPU time rather than invocation time, and allows a CPU to handle multiple requests.