@JorgeCastilloPr Maybe but Iβd rather kick something off or review it while Iβm on my way into the office so I can leave earlier without losing any time
@_KanyeDev Canβt agree more. Additional layers of abstraction just in the name of clean code and to then throw patterns at you to sound more authoritative. Avoid over indexing on those folks and rather step back and ask the question, βwhat would the customer want?β
@RepJackKimble Thatβs great, first part is already true. Any specialist is months if not years away from availability.
Why would communism result in beef shortages? And are you confusing that with socialism?
Yes we have over 400 subgraphs .
Fair points here, but a single field like node ids can propagate hundreds of times in a response, taking 40-45% of the entire response body if not carefully managed. Or adding a slow subgraph field without a defer can throw it off too.
For sure, having network integration tests can work well too
@nickmanurung@makuchaku@JorgeCastilloPr Yes and a single 1 minute method trace takes 8 GB of storage and takes more than 30 minutes to an hour to get a good answer from the agent on an M4 Pro
@makuchaku@JorgeCastilloPr Yeah we do. Except you get a lot variance in complex apps, that donβt always tell the full story compared to customer devices with all sorts of other apps and services
@JorgeCastilloPr Yes we can, however we still need some semblance of true numbers since changing GraphQL queries come with performance consequences too