𝗖𝗵𝗮𝘁𝘁𝘆 𝗔𝗣𝗜𝘀 𝗔𝗿𝗲 𝗞𝗶𝗹𝗹𝗶𝗻𝗴 𝗬𝗼𝘂𝗿 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲
Chatty APIs feel harmless.Until latency starts killing your system.
A common pattern:
Client → API call → API call → API call → API call
To build a single screen.
Each call looks small.
Together, they become expensive.
Problems with chatty APIs:
• High latency (too many round trips)
→ Every request takes time (network + processing). Multiple small calls add up and slow down the overall response
• Increased failure points
→ If one out of many calls fails, the whole user flow can break
• More load on network and services
→ Repeated calls increase traffic and put unnecessary pressure on backend systems
• Poor user experience (slow responses)
→ Users feel delays, especially on mobile or low network conditions
The issue isn’t just performance.
It’s design.
Better approach:
• Fetch more data in fewer calls (coarse-grained APIs)
→ Design APIs that return all required data in one response instead of forcing multiple calls
• Use aggregation at the backend (combine multiple calls into one)
→ Let the backend call different services internally and return a single combined response to the client
• Cache frequently used data
→ Store common responses so repeated requests don’t always hit the backend
• Avoid unnecessary network hops
→ Reduce service-to-service calls where possible; keep the path from request to response simple
Think in terms of:
“Give me everything I need”
instead of
“Let me ask for everything step by step”
Every network call has a cost.
Reduce the conversation.
Good APIs don’t talk more.
𝗧𝗵𝗲𝘆 𝘀𝗮𝘆 𝗺𝗼𝗿𝗲 𝗶𝗻 𝗳𝗲𝘄𝗲𝗿 𝘄𝗼𝗿𝗱𝘀.