Severe intermittent latency & 504s on Edge Functions in eu-central-1 - Anyone else experiencing this
edge functionsHey everyone, 
Iโm reaching out to see if anyone else hosted in the eu-central-1 (Frankfurt) region is experiencing massive, unexplainable latency spikes right now (and historically over the last few days).
What we are trying to achieve & What is going wrong:
Our Edge Functions, which normally execute in 200-300ms, are suddenly taking anywhere from 10 to 60+ seconds. Many of them are completely failing with 504 Gateway Timeout errors. During these spikes, database connections also seem to severely stall.
The weird part (Why we think it's a regional infrastructure issue):
We run the exact same software (identical DB schema, identical Edge Functions) across two completely separate Supabase projects in eu-central-1:
Project A: Pro Plan + Medium Compute (High traffic)
Project B: Free Plan + Nano Compute (Very low traffic)
The crazy thing is that both projects experience the exact same massive latency spikes at the exact same times. If this were just a poorly optimized query or memory overcommit on our end, the idle Free tier project wouldn't be crashing simultaneously with our active Pro project.
What we've tried / Context:
Support previously suggested a compute upgrade due to memory overcommit. We upgraded Project A to Medium. It didn't fix the issue immediately; the system magically stabilized a day later without any deployments on our end. Now, the exact same 504 issues are back.
We experienced this exact same "ghost" issue back in January. It eventually resolved itself after a few days without any intervention or code changes from us.
When inspecting the Edge Function logs, the functions just hang during the DB rpc calls until they timeout.
Has anyone else in the EU region noticed these sudden, simultaneous performance drops? Any tips on how to properly debug or prove whether this is a noisy neighbor / API gateway issue vs. something internal to our Postgres instances?
Thanks in advance for any insights!
Iโm reaching out to see if anyone else hosted in the eu-central-1 (Frankfurt) region is experiencing massive, unexplainable latency spikes right now (and historically over the last few days).
What we are trying to achieve & What is going wrong:
Our Edge Functions, which normally execute in 200-300ms, are suddenly taking anywhere from 10 to 60+ seconds. Many of them are completely failing with 504 Gateway Timeout errors. During these spikes, database connections also seem to severely stall.
The weird part (Why we think it's a regional infrastructure issue):
We run the exact same software (identical DB schema, identical Edge Functions) across two completely separate Supabase projects in eu-central-1:
Project A: Pro Plan + Medium Compute (High traffic)
Project B: Free Plan + Nano Compute (Very low traffic)
The crazy thing is that both projects experience the exact same massive latency spikes at the exact same times. If this were just a poorly optimized query or memory overcommit on our end, the idle Free tier project wouldn't be crashing simultaneously with our active Pro project.
What we've tried / Context:
Support previously suggested a compute upgrade due to memory overcommit. We upgraded Project A to Medium. It didn't fix the issue immediately; the system magically stabilized a day later without any deployments on our end. Now, the exact same 504 issues are back.
We experienced this exact same "ghost" issue back in January. It eventually resolved itself after a few days without any intervention or code changes from us.
When inspecting the Edge Function logs, the functions just hang during the DB rpc calls until they timeout.
Has anyone else in the EU region noticed these sudden, simultaneous performance drops? Any tips on how to properly debug or prove whether this is a noisy neighbor / API gateway issue vs. something internal to our Postgres instances?
Thanks in advance for any insights!
