Cloudflare workers randomly high wall time on New Zealand servers
Hi i would like help with a wall time issue that seems to randomly affect New Zealand traffic connections.
For context
- Our worker randomly experiences wall time in between Prisma Accelerate queries to Supabase for only New Zealand connections. - We handle 4x more visits from Australia and have visits from other countries that have never even once encountered this issue.
- One API call can branch into a Promise.all calling 4 functions asynchronously where each function can potentially make up to 3 Prisma queries.
- Issue can vary greatly, sometimes we can go for 4-5 hour periods where New Zealand connections get this issue every API call while other times we can go for 2-3 days with no problems
- Only affects our API's that branch off into many queries
As you can see in the images, there is a 30 second wall time (super consistent, its always around 30 seconds) that randomly applies. It usually always occurs at the exact same place but can also affect other things randomly. Once it occurs, it blocks everything, not just its own function but everything. Its like Cloudflare just stops for 30 seconds then starts working again.
Another thing to note - No error logs, no subquery errors. After 30 seconds, it continues running like nothing happened then returns everything and runs at normal speed (our typical wall time is 150-200ms, our queries are not long or complex and return little data).
Any help or pointers is appreciated as we are quite stuck as we think this may be a Cloudflare issue.
For context
- Our worker randomly experiences wall time in between Prisma Accelerate queries to Supabase for only New Zealand connections. - We handle 4x more visits from Australia and have visits from other countries that have never even once encountered this issue.
- One API call can branch into a Promise.all calling 4 functions asynchronously where each function can potentially make up to 3 Prisma queries.
- Issue can vary greatly, sometimes we can go for 4-5 hour periods where New Zealand connections get this issue every API call while other times we can go for 2-3 days with no problems
- Only affects our API's that branch off into many queries
As you can see in the images, there is a 30 second wall time (super consistent, its always around 30 seconds) that randomly applies. It usually always occurs at the exact same place but can also affect other things randomly. Once it occurs, it blocks everything, not just its own function but everything. Its like Cloudflare just stops for 30 seconds then starts working again.
Another thing to note - No error logs, no subquery errors. After 30 seconds, it continues running like nothing happened then returns everything and runs at normal speed (our typical wall time is 150-200ms, our queries are not long or complex and return little data).
Any help or pointers is appreciated as we are quite stuck as we think this may be a Cloudflare issue.
Welcome to the official Cloudflare Developers server. Here you can ask for help and stay updated with the latest news
86,942Members
Resources
Was this page helpful?
