Hitting the 1s data-tool limit of Sidekick App Extension

My Sidekick data extension’s handler runs in the Shopify sandbox, but the actual data lives on my own backend (a remote API + DB). So each tool call is: sandbox handler → fetch my backend → query/compute → return JSON, all within the ~1 second budget.

Some of my tools are analytics (aggregations over referrals/commissions) that aren’t instant to compute live.

  1. For those who’ve shipped Sidekick data extensions with a remote backend, does the network round-trip from the sandbox to your API realistically fit inside 1s in production?
  2. Is precompute + cache (background job storing results, handler just reads them) the standard approach, or are people computing live?
  3. Roughly what end-to-end latency have you seen from the sandbox, and does it vary by merchant region?
  4. When a tool occasionally exceeds 1s and gets skipped — is that per-call, or does Sidekick deprioritize the tool for a while afterward?

Trying to decide how much caching infrastructure to build up front. Thanks!

Hey @Ng_c_Nguy_n_Minh - thanks for reaching out.

As you mentioned, the documented response time for app data extensions is one second. At present, exceeding it doesn’t immediately cut off the response, and tools aren’t penalized or temporarily deprioritized for running over it. Shopify is still evaluating recurring response-time patterns before defining how consistently exceeding the limit might be handled in the future:

You should still aim to keep responses as close to the documented limit as possible. This avoids any future impact on your app, as well as offering the best UX to your users. Further guidance is available here: App tools

Hope this helps!