Platform Quality
Durable queues and webhooks
Reliable async processing for payments, comms, and integrations, work gets recorded and retried, not dropped under load.
Overview
When a payment, text, or integration event fires, it can't just vanish if a service hiccups. Thorbis is building durable queues and webhooks that record each event, process it reliably, and retry safely without double-charging or double-sending. The result is async work that holds up under real traffic, so payments clear and messages send even when volume spikes.
What we're building
When a payment, text, or integration event fires, it can't just vanish because a service hiccupped, a dropped event means a charge that never clears or a message that never sends.
Thorbis is building durable queues and webhooks that record each event, process it reliably, and retry safely without double-charging or double-sending. The work stays idempotent, so a duplicate webhook doesn't fire twice.
This is foundation work behind payments, communications, and integrations, async processing that holds up under real traffic so the day keeps moving even when volume spikes.
What you'll be able to do
- Record every payment, comms, and integration event durably
- Retry failed work safely without duplicates
- Process async work reliably under load
- Keep payments and messages from being dropped
- Stay idempotent against duplicate webhook events
- Hold up under traffic spikes without losing events
- Clear payments and send messages even when a service hiccups
Rollout plan
- 01
Durable event recording
In progressRecord every payment, comms, and integration event so nothing is lost.
- 02
Safe retries and idempotency
In progressRetry failed work without double-charging or double-sending.
- 03
Reliable under load
In progressKeep async processing steady through traffic spikes.
Milestones
- Not started:
Queue infrastructure in place
Stand up durable storage for async events.
- Not started:
Idempotent processing
Guard against duplicate webhook events.
- Not started:
Retry handling
Reprocess failed work safely.
- Not started:
Load validation
Confirm reliability under heavy volume.
More in Platform Quality
How to send useful feedback
Name the blocked role
Product feedback is more useful when it names the owner, dispatcher, technician, office user, or customer affected.
Explain the workaround
Describe what your team does today and where the process fails under real work.
Define success
Share what would change after launch: time saved, fewer manual steps, cleaner records, or better customer follow-up.