We’re seeing a serious latency regression on the Studio Memory API (GET api.pickaxe.co/v1/studio/memory/user/{id}, workspace studio-7c33e460). On July 5 reads measured 11-13 seconds each; today, July 7, the same reads measure 23-24 seconds, every call, even for a single filtered record. Small payloads, all time-to-first-byte, so it doesn’t look like a data-size issue on our end. We’ve built an app on this API and have had to raise timeouts twice in three days. Can you look at what changed, and is there a target response time we can design around?>>reads are now exceeding 40 seconds as of midday July 7.>> Fable says: That’s outage territory, and they should hear it while it’s happening, not after. If this is their new normal, my next build is server-side caching so actors never touch Pickaxe’s slow path directly, but let’s make them fix their pipes first.
We are seeing the same, we ping the API every 15 minutes to ensure things are up.
testing now. the agents are working, but API seems to be broken or struggling.
Hi @taedog2020 @actbold ,
Thank you for flagging this. We’re looking into it now and appreciate you reporting it.
We’ll share an update here once we have more information.