Webhooks
DevStride POSTs a signed JSON event to your HTTPS endpoint the moment an item, comment or SLA timer changes, so your systems react in seconds instead of polling the API.
On this page
Webhooks push changes out of DevStride as they happen. You register an HTTPS endpoint and choose the events it cares about; whenever one of those changes happens anywhere in your organization, DevStride sends your endpoint a small JSON document describing it, signed so you can prove it came from DevStride.
Instead of asking the API "has anything changed?" every few minutes, your system is told, usually within seconds, and only when there is something to tell.
Why webhooks instead of polling#
| Polling the API | Webhooks | |
|---|---|---|
| How fast you hear about a change | Your polling interval, at best | Usually seconds |
| Requests when nothing changes | One per interval, forever | None |
| Counts against your API rate limit | Every poll | Never: DevStride calls you |
| Knowing what changed | Diff the results yourself | Each event says what happened, who did it and how |
| Missed changes while your system is down | Depends on your diff logic | Queued and retried for about a day, then replayable for at least 30 days |
Webhooks and the API work well together: an event tells you that something changed and carries the essentials, and when you need the full record (a description, custom fields) you read it from the API by the ids in the event.
The events#
| Event | Sent when |
|---|---|
item.created |
A work item or Service Desk ticket is created, from any channel. |
item.status_changed |
An item moves to a different status. |
comment.created |
A comment is added to an item or ticket: public replies and internal notes alike. |
sla.at_risk |
A ticket is about to miss one of its SLA targets. |
sla.breached |
A ticket has missed one of its SLA targets. |
customer.message |
In headless mode, a message DevStride would have emailed a customer, for your own tool to send. |
The first three cover every work item in the organization; the two SLA events and customer.message concern Service Desk only. Every field of every event is listed in Events & Payloads.
What you can build#
- Sync tickets with a CRM or a partner's system. Mirror new tickets (
item.created), their progress (item.status_changed) and the conversation (comment.created) into Salesforce, HubSpot or an outsourced support desk, without a nightly export. - Chat bots for Slack, Microsoft Teams or Discord. Post to a channel when an urgent ticket arrives or an item reaches a Done status, with your own formatting and routing rules.
- Analytics and data pipelines. Stream every change into a warehouse (BigQuery, Snowflake, Redshift) or an event bus (Kafka, Amazon EventBridge) and measure cycle time to the second.
- Automation platforms. Start a Zapier, Make or n8n workflow from their "catch webhook" trigger. Verify the signature in a code step where the platform offers one, or put a small verifying relay in front.
- Audit trails. Keep an append-only record of who created and moved what, and when, in a system you control.
- SLA escalation and paging. Turn
sla.at_riskinto a Slack warning andsla.breachedinto a PagerDuty or Opsgenie incident, so a missed promise wakes the right person.
How it works#
- Something changes in DevStride: an item is created, a status changes, someone comments, an SLA timer crosses a threshold.
- DevStride builds one event for that change and queues one delivery for every enabled endpoint subscribed to that event type.
- Each delivery is POSTed to its endpoint as JSON, signed with that endpoint's own secret. Deliveries to one endpoint go out one at a time, in order.
- Your endpoint checks the signature, answers with any
2xxstatus, and processes the event. - If it does not answer
2xxin time, DevStride retries with growing waits for about a day. Everything is visible in the delivery log on the Webhooks settings page, where you can replay what failed.
The details are in Delivery, Retries & Ordering.
Webhooks or the "Call Webhook" automation action?#
DevStride also has a Call Webhook action in Automations. They solve different problems:
| Webhooks (this section) | Call Webhook automation action | |
|---|---|---|
| What triggers it | Every change of the subscribed types, organization-wide | One automation rule's trigger and conditions |
| Payload | A fixed, documented JSON contract | Text you compose from item fields |
| Signed | Yes, HMAC-SHA256 with a per-endpoint secret | No |
| Retries, ordering, delivery log, replay | Yes | No |
| Set up in | Configure Settings → Integrations → Webhooks | The automation builder |
Use webhooks to keep another system in sync. Use Call Webhook for a one-off notification shaped by a single rule.
Who can manage webhooks#
Everything on the Webhooks page, and every webhook API route, needs the Manage Webhooks permission (MANAGE_WEBHOOKS). The Owner and Admin roles hold it by default; you can grant it to a custom role in Roles & Permissions. It requires the Settings module. Members without it never see the page.
Grant Manage Webhooks with care
An endpoint receives events for every item and comment in the organization, internal notes included, regardless of what the person who set it up can see. For that reason the permission is not part of any permission template: applying the Manage org settings template to a role does not grant it. Give it only to people you would trust with an export of all your work.
Limits at a glance#
| Limit | Value |
|---|---|
| Endpoints per organization | 20 |
| Endpoint URL | https:// only, at most 2,048 characters; private, loopback, link-local and cloud-metadata addresses are refused |
| Time to answer | 10 seconds per request, 15 seconds in total including DNS lookups and redirects |
| Response body DevStride reads | 64 KB (a larger response counts as a failure) |
| Delivery attempts | 10, spread over about 24 hours |
| Signature timestamp tolerance | 5 minutes (300 seconds) |
| Previous secret still valid after a rotation | 24 hours |
Comment HTML carried in comment.created |
256 KB (larger comments arrive with truncated: true) |
| Delivery log retention | At least 30 days after a delivery was queued |
Where to go next#
- Quickstart: a working, verified receiver in about ten minutes.
- Events & Payloads: the envelope and every field of every event, with examples.
- Verifying Signatures: the signing scheme and tested code in Node.js, Python, Go, Ruby, PHP and C#.
- Delivery, Retries & Ordering: what counts as success, the retry schedule, ordering and health alerts.
- Managing Endpoints: editing, disabling, rotating secrets, the delivery log and replay, from the app, the REST API or the MCP server.
- Troubleshooting & Best Practices: building a receiver that never loses or double-counts an event.