Limits and quotas
In addition to the account-level limits described on the
pricing page, the Airplane API and runtime have a few
lower-level limits to keep performance high and reduce the potential for abuse. See the table below
for a summary of these.
Limit/quota summary
Quantity | Maximum value | Notes |
|---|---|---|
| Task run timeout | 12 hours (standard) 60 days (workflow) | See the task runtimes page for more information. |
| CPU per Airplane-hosted task run | 1 vCPU | Runs managed by self-hosted agents can have arbitrarily high CPU limits. |
| Memory per Airplane-hosted task run | 4GB | Runs managed by self-hosted agents can have arbitrarily high memory limits. |
| Concurrent executions per task | 10 (free accounts) 25 (paid accounts) | |
| Total pending executions per team | 2500 | |
| Log lines per task run | 1 million | |
| Total log size per task run | 200MB | |
| Total output size per task run | 150MB |
If you have any questions about these or believe they will impede your use of the product, please
let us know!
Handling limits
The Airplane API will return a
429 error if a request has caused an
API-based limit to be exceeded. The client should then try again later, preferably with an
exponential backoff.