Increase in Accelerate cold start time
Incident Lifecycle
Incident Timeline
Monitoring
A fix has been implemented and we are monitoring the results.
Apr 2, 2026 at 5:43 AM UTC
Identified
We are awaiting an update from Cloudflare.
As the incident persists, we recommend not deploying new versions of your schema until the incident is resolved to avoid requesting a new Query Engine and encountering a cold start.
Workarounds: For Prisma Postgres users we still recommend Direct TCP with connection pooling enabled.
As the incident persists, we recommend not deploying new versions of your schema until the incident is resolved to avoid requesting a new Query Engine and encountering a cold start.
Workarounds: For Prisma Postgres users we still recommend Direct TCP with connection pooling enabled.
Apr 2, 2026 at 2:56 AM UTC
Identified
We've isolated the root cause to an issue within Durable Objects. Cloudflare have an active status page.
We're investigating mitigation options.
We're investigating mitigation options.
Apr 2, 2026 at 12:57 AM UTC
Investigating
We are continuing to investigate the issue.
Affected users: US East, AP Northeast, AP Southeast, US West, EU Central
Root cause: Currently unknown
Workarounds: For Prisma Postgres users we recommend Direct TCP with connection pooling enabled.
Affected users: US East, AP Northeast, AP Southeast, US West, EU Central
Root cause: Currently unknown
Workarounds: For Prisma Postgres users we recommend Direct TCP with connection pooling enabled.
Apr 1, 2026 at 11:36 PM UTC
Investigating
Description: We are investigating an issue where Accelerate cold starts have increased.
Affected users: US East
Root cause: Currently unknown
Workarounds: For Prisma Postgres users we recommend Direct TCP with connection pooling enabled.
Affected users: US East
Root cause: Currently unknown
Workarounds: For Prisma Postgres users we recommend Direct TCP with connection pooling enabled.
Apr 1, 2026 at 10:15 PM UTC
Was your business affected by this Prisma outage?
Set up instant alerts for Prisma — be the first to know about outages via email, Slack, Teams, or Discord.