Rate limits and retention
Rate limits
Tillcast does not limit the number of requests per key today. The limits that exist protect a printer and the service from loops:
A limited request gets HTTP 429. Wait for the number of seconds in Retry-After, then retry (queue_full sends 30). A queue empties as the printer prints, as you cancel jobs and as jobs pass their expires_at (24 hours by default), so do not retry queue_full in a loop. The SDKs wait for Retry-After inside the call, so one call can block for about a minute: see Retry safely. Handle 429 in your client now, so that a limit added later does not surprise you.
Be kind to the service even without a limit:
- Lists that span a whole organization, such as all printers, are slower than a list by location or by customer. Prefer a list by location, and do not sweep large lists every few seconds.
- Do not poll for job state faster than you need. A printer takes a job within about one poll interval, 5 seconds by default. Webhooks, which avoid polling, are planned and not built. See Webhooks and events.
Job limits
See Jobs and Receipt documents.
Retention
Some data does not exist yet, because the features that make it are planned: printer state events (3 months), lifecycle events (1 year), the webhook delivery log (30 days), and usage statistics. Their periods are decided. The full table is in Security and data. There is no option to keep data longer, not even for a fee.