Limits

These are the current repository defaults. Production configuration can override the runtime values.

Event execution

Limit	Default
Wall-clock time for one event	10 seconds
Active V8 CPU time for one event	500 milliseconds
Discord REST requests per event	20
KV operations per event	100
fetch() timeout	10 seconds
Maximum fetch() response size	10 MiB

Waiting for Discord, KV, or fetch() consumes wall-clock time but not active CPU time. Keep event handlers short and move scheduled work into tasks.

Compute jobs

Limit	Default
CPU time requested by one compute job	1 second
Compute CPU bucket	10 seconds per event
Compute jobs started by one event	8

weeble.compute.run() uses this separate CPU bucket for bounded CPU-heavy work. It does not create a durable background job and must still finish within its wall-clock timeout.

Builds and revisions

Limit	Default
Build timeout	30 seconds
Build request size	3 MiB
Recent revisions retained per deployment	25

Active, current, observed, desired, and scheduled-task revisions may be retained in addition to the 25-revision window.

npm package review

Limit	Value
Packages in one resolved dependency graph	100
npm metadata for one lookup	2 MiB
Package archive	2 MiB
Extracted supported files	5 MiB
Supported files in one package	500
New package requests	5 per minute

Every requested root package uses an exact x.y.z version. Transitive dependencies are resolved and approved as exact immutable artifacts.

Account entitlements

Entitlement	Free	Pro
Accessible guilds	3	10
Deployments	10	100
Custom bot identities	0	1
Runtime log retention	3 days	30 days

Pro access is granted for a limited period by the Weeble team.

Note

Discord also applies its own endpoint and global rate limits. Weeble coordinates those limits, but code should still avoid unnecessary request loops.