DO resets at ~30s despite cpu_ms=300000

S

SkoogZ•8/26/25, 8:26 AM

Hi team — I’ve been trying to run a Durable Object that does a bunch of awaited fetches and then a CPU-heavier finalize step, but the invocation keeps resetting at ~30s even though I’ve set a 5-minute CPU budget.

Context

Plan: Paid • Worker: Durable Object
Dashboard shows CPU time limit: 300000 ms

Deployed (not

wrangler dev

wrangler dev

) •

nodejs_compat

nodejs_compat

on

Config (sanitized)

name = "redacted-worker"
main = "src/index.ts"
compatibility_flags = [ "nodejs_compat" ]
compatibility_date = "2025-06-10"

[durable_objects]
bindings = [{ name = "MY_DO", class_name = "MyDO" }]

[limits]
cpu_ms = 300000

name = "redacted-worker"
main = "src/index.ts"
compatibility_flags = [ "nodejs_compat" ]
compatibility_date = "2025-06-10"

[durable_objects]
bindings = [{ name = "MY_DO", class_name = "MyDO" }]

[limits]
cpu_ms = 300000

name = "redacted-worker"
main = "src/index.ts"
compatibility_flags = [ "nodejs_compat" ]
compatibility_date = "2025-06-10"

[durable_objects]
bindings = [{ name = "MY_DO", class_name = "MyDO" }]

[limits]
cpu_ms = 300000

name = "redacted-worker"
main = "src/index.ts"
compatibility_flags = [ "nodejs_compat" ]
compatibility_date = "2025-06-10"

[durable_objects]
bindings = [{ name = "MY_DO", class_name = "MyDO" }]

[limits]
cpu_ms = 300000

Symptom

Invocation resets at ~30s with:
```
durable object exceeded its cpu time limit and was reset
```
durable object exceeded its cpu time limit and was reset
```
durable object exceeded its cpu time limit and was reset
```
durable object exceeded its cpu time limit and was reset
.
Handler: many awaited network calls, then a CPU-heavier finalize step (encode/compress + single upload).
No
```
state.blockConcurrencyWhile
```
state.blockConcurrencyWhile
```
state.blockConcurrencyWhile
```
state.blockConcurrencyWhile
or
```
state.storage.transaction
```
state.storage.transaction
```
state.storage.transaction
```
state.storage.transaction
in this path.

What I tried

Redeployed after adding
```
[limits]
```
[limits]
```
[limits]
```
[limits]
.
Reset occurs only on the finalize/compression path.

Happy to DM a project privately if needed. Thanks!

N

Nicola•8/26/25, 9:54 AM

Hey, we encounter the same issue.
We configured it the same way as you did, and in the Cloudflare dashboard, it already shows up as 300_000 ms.

I suspect that maybe the durable object worker has 5 minutes, but the durable object itself has less.
In our implementation, calls from other workers are directly sent to the durable object and never reach the worker's fetch handler.
I don't know how to confirm if that's the case, but that's what I think could be the problem.

NNicola Hey, we encounter the same issue. We configured it the same way as you did, and...

S

SkoogZ•8/26/25, 11:10 AM

In my case, I'm invoking the DO directly from the worker that it is defined in. So I guess the 30s applies regardless.

Also, based on the documentation stating that the CPU time limit for the DO can be increased to up to 5 minutes, I don't see it only applying to the DO's worker.

DO resets at ~30s despite cpu_ms=300000

Similar Threads

Similar Threads

Similar Threads