Incident Summary A health check designed to prevent Out-of-Memory (OOM) conditions began closing some incoming connections, which led to 503 errors for a subset of users. These errors were initially surfaced under a generic SUPABASE_EDGE_RUNTIME_ERROR code.
Timeline & Actions Sunday, 26: Issue detected as user reports of 503 errors increased. Investigation into root cause began. Monday, 27: Introduced a more accurate error code (SUPABASE_EDGE_RUNTIME_SERVICE_DEGRADED) and implemented infrastructure changes to mitigate impact. Tuesday, 28: Adjusted conditions for returning 503 responses to be less aggressive. Prepared a “retry-on-degraded” mechanism (not yet deployed).
Current Status Mitigations are in place and improvements to error handling are ongoing. Further resilience enhancements will be deployed shortly.
No components marked as affected
Resolved
Incident Summary A health check designed to prevent Out-of-Memory (OOM) conditions began closing some incoming connections, which led to 503 errors for a subset of users. These errors were initially surfaced under a generic SUPABASE_EDGE_RUNTIME_ERROR code.
Timeline & Actions Sunday, 26: Issue detected as user reports of 503 errors increased. Investigation into root cause began. Monday, 27: Introduced a more accurate error code (SUPABASE_EDGE_RUNTIME_SERVICE_DEGRADED) and implemented infrastructure changes to mitigate impact. Tuesday, 28: Adjusted conditions for returning 503 responses to be less aggressive. Prepared a “retry-on-degraded” mechanism (not yet deployed).
Current Status Mitigations are in place and improvements to error handling are ongoing. Further resilience enhancements will be deployed shortly.