Events - Runcrate

Every event arrives in the same envelope. This page lists each type, when it is sent and what its data holds. The examples are the exact JSON a signed webhook endpoint receives. Slack and Discord get the same event drawn as a message (see Slack and Discord).

Versioning

Each event carries api_version, a date naming the format it follows. The examples below show the current one. Within one version:

data is open. New keys can appear in any event at any time, so ignore keys you do not use and never reject an event for having one you do not know.
An existing key never changes its meaning or its type.
New event types can be added. An endpoint set to All events receives them; one set to Choose events only gets what it picked.

A change that would break a receiver comes with a new api_version.

Choosing events

In Settings > Webhooks, each endpoint has an Events setting: All events, or Choose events to pick from the groups below. webhook.test is the exception: it goes to the endpoint you test, whatever it subscribes to. Fields are snake_case. Money is in USD as a number (412.37), times are ISO 8601 in UTC, and ids are UUIDs. In instance events, deployment_id is the instance id you get from the API, CLI and MCP tools.

Event	Group	What it means
`budget.threshold_reached`	Budgets	An environment’s month to date reached 80% of its monthly budget.
`budget.reached`	Budgets	An environment spent its full monthly budget: launches are refused, instances stop and its API keys are declined.
`instance.ready`	Instances	An instance or crate is running and reachable over SSH and its ports.
`instance.launch_failed`	Instances	A launch did not start. Nothing was charged for it.
`instance.runtime_ending`	Instances	An instance is 10 minutes from its max runtime.
`instance.auto_stopped`	Instances	Runcrate stopped an instance: max runtime, idle stop, a budget or a low balance.
`instance.stopped`	Instances	An instance was terminated for any reason. stop_reason says why.
`job.queued`	Jobs	A job is waiting: no capacity for the GPUs it asked for yet, or the budget. Sent once per attempt.
`job.started`	Jobs	A job’s command started running on its machine.
`job.finished`	Jobs	A job ended: succeeded, failed, timed out or cancelled.
`schedule.failed`	Schedules	A scheduled launch did not start.
`schedule.paused`	Schedules	A schedule was turned off: 5 scheduled runs in a row failed, or who it runs as lost access.
`balance.low`	Billing	The balance is running low. When stops_at is set, running instances stop then unless money is added.
`balance.depleted`	Billing	The balance ran out and running instances were stopped.
`credits.added`	Billing	A payment or top-up was added to the workspace balance.
`storage.suspended`	Storage	A storage volume was suspended because its storage could not be paid for.
`storage.reactivated`	Storage	A suspended storage volume is available again.
`webhook.test`	Webhooks	Sent by Send a test event. Never sent on its own.
`webhook.disabled`	Webhooks	An endpoint kept failing (3 days with no success and at least 10 failed attempts, or 20 events in a row given up on), so Runcrate turned it off.

Budgets

budget.threshold_reached

An environment’s month to date reached 80% of its monthly budget. When: Once per environment, month and budget amount, when the month to date first reaches 80% of the budget. Raising the budget can send it again at the new amount. Owners and admins also get an email.

Field	Type	Meaning
`environment_id`	string (UUID)	The environment the budget belongs to.
`environment_name`	string	The environment’s name.
`monthly_budget_usd`	number	The environment’s monthly budget, in USD.
`month`	string	The UTC month the spend counts toward, as `YYYY-MM`.
`month_to_date_usd`	number	What the environment has spent this month (GPU instances and inference), in USD.
`percent`	integer	How much of the budget is spent, in percent: `80`.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9000",
  "type": "budget.threshold_reached",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
    "environment_name": "production",
    "monthly_budget_usd": 500,
    "month": "2026-10",
    "month_to_date_usd": 402.17,
    "percent": 80
  }
}

budget.reached

An environment spent its full monthly budget: launches are refused, instances stop and its API keys are declined. When: Once per environment, month and budget amount, when the month to date reaches the budget. From then on launches in the environment fail with HTTP 402, its running instances are terminated (each sends instance.auto_stopped with stop_reason budget) and its API keys are declined, until the next UTC month or until the budget is raised. Owners and admins also get an email.

Field	Type	Meaning
`environment_id`	string (UUID)	The environment the budget belongs to.
`environment_name`	string	The environment’s name.
`monthly_budget_usd`	number	The environment’s monthly budget, in USD.
`month`	string	The UTC month the spend counts toward, as `YYYY-MM`.
`month_to_date_usd`	number	What the environment has spent this month (GPU instances and inference), in USD.
`percent`	integer	How much of the budget is spent, in percent: `100`.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9001",
  "type": "budget.reached",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
    "environment_name": "production",
    "monthly_budget_usd": 500,
    "month": "2026-10",
    "month_to_date_usd": 500.42,
    "percent": 100
  }
}

Instances

instance.ready

An instance or crate is running and reachable over SSH and its ports. When: Once per launch, when the instance or crate reaches running.

Field	Type	Meaning
`deployment_id`	string (UUID)	The instance’s `id`, the same one the API, CLI and MCP tools use.
`deployment_name`	string or `null`	The instance’s name.
`environment_id`	string (UUID)	The environment the instance runs in.
`kind`	string	`instance`, or `crate` for a crate launch.
`gpu_type`	string or `null`	The GPU model, for example `H100`.
`gpu_count`	integer or `null`	How many GPUs it has.
`region`	string or `null`	The region it runs in.
`ip`	string or `null`	Its public IP address. `null` when it has none yet.
`ssh_command`	string or `null`	The command that connects to it over SSH. `null` when it has no IP yet.
`hourly_rate_usd`	number	What it costs per hour, in USD.
`ready_at`	string (ISO 8601)	When it became ready.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9002",
  "type": "instance.ready",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "deployment_id": "7d2e9c41-5b3a-4f60-8e17-c4a0b9d2f655",
    "deployment_name": "train-llama-8b",
    "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
    "kind": "instance",
    "gpu_type": "H100",
    "gpu_count": 1,
    "region": "us-east",
    "ip": "203.0.113.24",
    "ssh_command": "ssh root@203.0.113.24",
    "hourly_rate_usd": 2.49,
    "ready_at": "2026-10-20T14:02:11.000Z"
  }
}

instance.launch_failed

A launch did not start. Nothing was charged for it. When: When a launch ends before the instance ever ran. A launch that never ran is not billed.

Field	Type	Meaning
`deployment_id`	string (UUID)	The instance’s `id`, the same one the API, CLI and MCP tools use.
`deployment_name`	string or `null`	The instance’s name.
`environment_id`	string (UUID)	The environment the instance runs in.
`failure_reason`	string	A stable code for the failure: `create_failed`, `machine_error` or `provision_timeout`. Show `message` to people, not this code.
`message`	string	One sentence about what happened, fit to show to people.
`failed_at`	string (ISO 8601)	When the launch failed.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9003",
  "type": "instance.launch_failed",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "deployment_id": "7d2e9c41-5b3a-4f60-8e17-c4a0b9d2f655",
    "deployment_name": "train-llama-8b",
    "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
    "failure_reason": "provision_timeout",
    "message": "The machine took too long to start, so it was cancelled. You were not charged.",
    "failed_at": "2026-10-20T14:02:11.000Z"
  }
}

instance.runtime_ending

An instance is 10 minutes from its max runtime. When: Ten minutes before an instance reaches its max runtime. Changing the max runtime sends a fresh one ten minutes before the new stop time. Owners and admins also get an email.

Field	Type	Meaning
`deployment_id`	string (UUID)	The instance’s `id`, the same one the API, CLI and MCP tools use.
`deployment_name`	string	The instance’s name.
`environment_id`	string (UUID)	The environment the instance runs in.
`auto_stop_at`	string (ISO 8601)	When the instance will be terminated.
`max_runtime_minutes`	integer	The max runtime set on the instance, in minutes.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9004",
  "type": "instance.runtime_ending",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "deployment_id": "7d2e9c41-5b3a-4f60-8e17-c4a0b9d2f655",
    "deployment_name": "train-llama-8b",
    "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
    "auto_stop_at": "2026-10-20T14:12:11.000Z",
    "max_runtime_minutes": 240
  }
}

instance.auto_stopped

Runcrate stopped an instance: max runtime, idle stop, a budget or a low balance. When: When Runcrate terminates an instance on its own: stop_reason is max_runtime, idle, budget or low_balance. The same termination also sends instance.stopped. Owners and admins also get an email.

Field	Type	Meaning
`deployment_id`	string (UUID)	The instance’s `id`, the same one the API, CLI and MCP tools use.
`deployment_name`	string	The instance’s name.
`environment_id`	string (UUID)	The environment the instance runs in.
`stop_reason`	string	Why it was stopped: `max_runtime`, `idle`, `budget` or `low_balance`.
`stopped_at`	string (ISO 8601)	When it was stopped.
`max_runtime_minutes`	integer or `null`	The instance’s max runtime in minutes, or `null` when it had none.
`idle_shutdown_minutes`	integer or `null`	The idle stop in minutes when idle stop stopped the instance, otherwise `null`.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9005",
  "type": "instance.auto_stopped",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "deployment_id": "7d2e9c41-5b3a-4f60-8e17-c4a0b9d2f655",
    "deployment_name": "train-llama-8b",
    "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
    "stop_reason": "max_runtime",
    "stopped_at": "2026-10-20T14:02:11.000Z",
    "max_runtime_minutes": 240,
    "idle_shutdown_minutes": null
  }
}

instance.stopped

An instance was terminated for any reason. stop_reason says why. When: Every time an instance is terminated, whoever or whatever stopped it.

Field	Type	Meaning
`deployment_id`	string (UUID)	The instance’s `id`, the same one the API, CLI and MCP tools use.
`deployment_name`	string	The instance’s name.
`environment_id`	string (UUID)	The environment the instance runs in.
`stop_reason`	string	Why it stopped: `user`, `max_runtime`, `idle`, `budget`, `low_balance`, `failed`, `system` or `job_finished`.
`stopped_at`	string (ISO 8601)	When it stopped.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9006",
  "type": "instance.stopped",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "deployment_id": "7d2e9c41-5b3a-4f60-8e17-c4a0b9d2f655",
    "deployment_name": "train-llama-8b",
    "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
    "stop_reason": "user",
    "stopped_at": "2026-10-20T14:02:11.000Z"
  }
}

Jobs

Every job event also carries attempt (1-based), attempts_count, group_id (null for a single job) and tags. A job’s own webhook_url gets its job.started and job.finished too, signed the same way (see Jobs: idempotency and callbacks).

job.queued

A job is waiting: no GPU it asked for has capacity yet, or its environment’s budget is spent. When: Once per attempt, when the job starts waiting. It is tried again until its capacity_wait_minutes are over, then it fails.

Field	Type	Meaning
`job_id`	string (UUID)	The job’s `id`.
`job_name`	string	The job’s name.
`environment_id`	string (UUID)	The environment the job runs in.
`reason`	string	`no_capacity` or `budget`.
`gpu_candidates`	array	The GPUs it asked for, in order, as `{ "gpu_type", "gpu_count" }`.
`next_attempt_at`	string (ISO 8601)	When it is tried next.
`wait_until`	string (ISO 8601)	When it stops waiting and fails.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9006",
  "type": "job.queued",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "job_id": "0b6f3e28-9c1d-4a75-b2e4-81f7c3d5a960",
    "job_name": "nightly-eval",
    "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
    "reason": "no_capacity",
    "gpu_candidates": [{ "gpu_type": "H100", "gpu_count": 1 }, { "gpu_type": "A100-80GB", "gpu_count": 2 }],
    "next_attempt_at": "2026-10-20T14:03:11.000Z",
    "wait_until": "2026-10-20T14:32:11.000Z",
    "attempt": 1,
    "attempts_count": 1,
    "group_id": null,
    "tags": { "team": "ml" }
  }
}

job.started

A job’s command started running on its machine. When: When a job’s machine is ready and its command starts: once per attempt.

Field	Type	Meaning
`job_id`	string (UUID)	The job’s `id`.
`job_name`	string	The job’s name.
`environment_id`	string (UUID)	The environment the job runs in.
`gpu_type`	string	The GPU model, for example `A100`.
`gpu_count`	integer	How many GPUs the job has.
`max_runtime_minutes`	integer	The job’s maximum runtime, in minutes.
`started_at`	string (ISO 8601)	When the command started.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9007",
  "type": "job.started",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "job_id": "0b6f3e28-9c1d-4a75-b2e4-81f7c3d5a960",
    "job_name": "nightly-eval",
    "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
    "gpu_type": "A100",
    "gpu_count": 1,
    "max_runtime_minutes": 120,
    "started_at": "2026-10-20T14:02:11.000Z",
    "attempt": 1,
    "attempts_count": 1,
    "group_id": null,
    "tags": { "team": "ml" }
  }
}

job.finished

A job ended: succeeded, failed, timed out or cancelled. When: Once per job, when it reaches a final status (a retried attempt does not end the job). A job whose machine never started gets one too, with status failed or cancelled and null for what never happened.

Field	Type	Meaning
`job_id`	string (UUID)	The job’s `id`.
`job_name`	string	The job’s name.
`environment_id`	string (UUID)	The environment the job runs in.
`status`	string	`succeeded`, `failed`, `timed_out` or `cancelled`.
`exit_code`	integer or `null`	The command’s exit code. `null` when the command never reported one.
`error`	string or `null`	One sentence about why the job did not succeed. `null` when it succeeded.
`started_at`	string (ISO 8601) or `null`	When the command started. `null` when it never started.
`ended_at`	string (ISO 8601)	When the job ended.
`duration_seconds`	integer or `null`	How long the command ran, in seconds. `null` when it never started.
`cost_usd`	number or `null`	What the job’s machines cost, every attempt added up, in USD. `null` when no machine ran.
`logs_url`	string	The API URL of the job’s logs. Call it with an API key.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9008",
  "type": "job.finished",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "job_id": "0b6f3e28-9c1d-4a75-b2e4-81f7c3d5a960",
    "job_name": "nightly-eval",
    "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
    "status": "succeeded",
    "exit_code": 0,
    "error": null,
    "started_at": "2026-10-20T13:15:11.000Z",
    "ended_at": "2026-10-20T14:02:11.000Z",
    "duration_seconds": 2820,
    "cost_usd": 1.37,
    "logs_url": "https://www.runcrate.ai/api/v1/jobs/0b6f3e28-9c1d-4a75-b2e4-81f7c3d5a960/logs",
    "attempt": 1,
    "attempts_count": 1,
    "group_id": null,
    "tags": { "team": "ml" }
  }
}

Schedules

schedule.failed

A scheduled launch did not start. When: When a scheduled run was due and did not start. Nothing is charged for it. When the failure is the 5th scheduled run in a row, the schedule is paused and schedule.paused is sent instead. See schedules. Owners and admins also get an email.

Field	Type	Meaning
`schedule_id`	string (UUID)	The schedule’s `id`.
`schedule_name`	string	The schedule’s name.
`kind`	string	What the schedule launches: `instance` or `job`.
`error`	string	One sentence about why the launch did not start.
`scheduled_for`	string (ISO 8601)	When the run was due.
`recurring`	boolean	Whether the schedule repeats. A one-off schedule does not run again.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9009",
  "type": "schedule.failed",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "schedule_id": "5e8a1c73-2f4b-4d09-9a6e-3b7c0d2f1e48",
    "schedule_name": "weekday-dev-box",
    "kind": "instance",
    "error": "No machine of this type was available. Nothing was charged.",
    "scheduled_for": "2026-10-20T14:02:11.000Z",
    "recurring": true
  }
}

schedule.paused

A schedule was turned off: 5 scheduled runs in a row failed, or who it runs as lost access. When: When Runcrate turns a schedule off: 5 scheduled runs in a row failed, the person it runs as left the workspace, or the API key or agent sign-in behind it was revoked or ended. It does not run again until someone resumes it. Owners and admins also get an email.

Field	Type	Meaning
`schedule_id`	string (UUID)	The schedule’s `id`.
`schedule_name`	string	The schedule’s name.
`kind`	string	What the schedule launches: `instance` or `job`.
`reason`	string	`failing` (5 scheduled runs in a row failed), `no_access` (the person it runs as left the workspace) or `credential_revoked` (its API key or agent sign-in was revoked or ended).
`consecutive_failures`	integer or `null`	How many scheduled runs in a row failed. `null` unless `reason` is `failing`.
`error`	string	The last failure, as one sentence.
`paused_at`	string (ISO 8601)	When the schedule was paused.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9010",
  "type": "schedule.paused",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": "e81b4f6a-2c3d-4a9e-b710-5f8c2d1e9a04",
  "data": {
    "schedule_id": "5e8a1c73-2f4b-4d09-9a6e-3b7c0d2f1e48",
    "schedule_name": "weekday-dev-box",
    "kind": "instance",
    "reason": "failing",
    "consecutive_failures": 5,
    "error": "No machine of this type was available. Nothing was charged.",
    "paused_at": "2026-10-20T14:02:11.000Z"
  }
}

Billing

balance.low

The balance is running low. When stops_at is set, running instances stop then unless money is added. When: When the workspace balance falls under $10, and again when it falls under $1 and a grace period starts: running instances stop at stops_at unless funds arrive. Add funds or turn on auto-recharge to keep instances running. Workspace-wide: the envelope’s environment_id is null.

Field	Type	Meaning
`balance_usd`	number	The workspace balance, in USD.
`stops_at`	string (ISO 8601) or `null`	When running instances will stop unless funds are added. `null` for the under-$10 warning.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9011",
  "type": "balance.low",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": null,
  "data": {
    "balance_usd": 4.12,
    "stops_at": null
  }
}

balance.depleted

The balance ran out and running instances were stopped. When: When the workspace balance runs out. Running instances are terminated, each with stop_reason low_balance. Workspace-wide: the envelope’s environment_id is null.

Field	Type	Meaning
`balance_usd`	number	The workspace balance, in USD. It can be below zero.
`instances_stopped`	integer	How many running instances were stopped.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9012",
  "type": "balance.depleted",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": null,
  "data": {
    "balance_usd": 0,
    "instances_stopped": 2
  }
}

credits.added

A payment or top-up was added to the workspace balance. When: When money is added to the workspace balance: a top-up, an auto-recharge or an adjustment. Workspace-wide: the envelope’s environment_id is null.

Field	Type	Meaning
`amount_usd`	number	The amount added, in USD.
`balance_usd`	number	The workspace balance after it was added, in USD.
`source`	string	`top_up`, `auto_recharge` or `adjustment`.
`transaction_id`	string (UUID)	The `id` of the transaction on the billing page.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9013",
  "type": "credits.added",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": null,
  "data": {
    "amount_usd": 100,
    "balance_usd": 104.12,
    "source": "top_up",
    "transaction_id": "b7e2c9a1-3d5f-4e08-9c6b-1a2f4d8e0c37"
  }
}

Storage

storage.suspended

A storage volume was suspended because its storage could not be paid for. When: When a storage volume is suspended because the workspace could not pay for it. storage.reactivated follows once it is paid for. Workspace-wide: the envelope’s environment_id is null.

Field	Type	Meaning
`storage_id`	string (UUID)	The volume’s `id`.
`storage_name`	string	The volume’s name.
`size_gb`	integer	The volume’s size, in GB.
`reason`	string	Why it was suspended: `unpaid`.
`suspended_at`	string (ISO 8601)	When it was suspended.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9014",
  "type": "storage.suspended",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": null,
  "data": {
    "storage_id": "9c4d7b12-6e0a-4f38-a5c1-2d8e7f3b0a69",
    "storage_name": "datasets",
    "size_gb": 500,
    "reason": "unpaid",
    "suspended_at": "2026-10-20T14:02:11.000Z"
  }
}

storage.reactivated

A suspended storage volume is available again. When: When a suspended volume is paid for and can be attached again. Workspace-wide: the envelope’s environment_id is null.

Field	Type	Meaning
`storage_id`	string (UUID)	The volume’s `id`.
`storage_name`	string	The volume’s name.
`size_gb`	integer	The volume’s size, in GB.
`reactivated_at`	string (ISO 8601)	When it became available again.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9015",
  "type": "storage.reactivated",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": null,
  "data": {
    "storage_id": "9c4d7b12-6e0a-4f38-a5c1-2d8e7f3b0a69",
    "storage_name": "datasets",
    "size_gb": 500,
    "reactivated_at": "2026-10-20T14:02:11.000Z"
  }
}

Webhooks

webhook.test

Sent by Send a test event. Never sent on its own. When: Only when someone presses Send a test event, and only to that one endpoint, whatever events it subscribes to. You cannot subscribe to it. Each person can send 10 tests per 10 minutes. Workspace-wide: the envelope’s environment_id is null.

Field	Type	Meaning
`test`	boolean	Always `true`.
`message`	string	A sentence saying this is a test.
`webhook_id`	string (UUID)	The endpoint the test was sent to.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9016",
  "type": "webhook.test",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": null,
  "data": {
    "test": true,
    "message": "A test event from Runcrate. Nothing happened in your workspace.",
    "webhook_id": "d1f0a7c3-4e2b-4c86-9b5d-7a3e1f0c8b24"
  }
}

webhook.disabled

An endpoint kept failing (3 days with no success and at least 10 failed attempts, or 20 events in a row given up on), so Runcrate turned it off. When: When Runcrate turns an endpoint off (see automatic turn-off). It goes to your other endpoints that subscribe to it, never to the one turned off. Owners and admins also get an email. Workspace-wide: the envelope’s environment_id is null.

Field	Type	Meaning
`webhook_id`	string (UUID)	The endpoint that was turned off.
`channel`	string	`webhook`, `slack` or `discord`.
`url_display`	string	The endpoint’s host and the last four characters of its URL. The full URL is never sent.
`reason`	string	Why it was turned off: `failing`.
`failing_since`	string (ISO 8601)	When the run of failures began.
`last_error`	string	The last failure, as one sentence.

Example

{
  "id": "6f1d3c2a-8a4e-4b7f-9a51-2f0c1e7d9017",
  "type": "webhook.disabled",
  "api_version": "2026-10-20",
  "created_at": "2026-10-20T14:02:11.000Z",
  "workspace_id": "a3c9e2f0-1b7d-4e55-8f20-6d4b9c1a7e33",
  "environment_id": null,
  "data": {
    "webhook_id": "d1f0a7c3-4e2b-4c86-9b5d-7a3e1f0c8b24",
    "channel": "webhook",
    "url_display": "hooks.example.com/…f3a9",
    "reason": "failing",
    "failing_since": "2026-10-17T14:02:11.000Z",
    "last_error": "The server behind this URL had an error (it answered 502)."
  }
}

Verifying signatures

Check that an event came from Runcrate.

Retries and the delivery log

What happens when your endpoint does not answer.

Spend guardrails

Max runtime, idle stop and monthly budgets, which send the budget and auto stop events.

Best practices

Dedupe, ordering and keeping the secret safe.

​Versioning

​Choosing events

​Budgets

​budget.threshold_reached

​budget.reached

​Instances

​instance.ready

​instance.launch_failed

​instance.runtime_ending

​instance.auto_stopped

​instance.stopped

​Jobs

​job.queued

​job.started

​job.finished

​Schedules

​schedule.failed

​schedule.paused

​Billing

​balance.low

​balance.depleted

​credits.added

​Storage

​storage.suspended

​storage.reactivated

​Webhooks

​webhook.test

​webhook.disabled

​Related

Verifying signatures

Retries and the delivery log

Spend guardrails

Best practices

Versioning

Choosing events

Budgets

budget.threshold_reached

budget.reached

Instances

instance.ready

instance.launch_failed

instance.runtime_ending

instance.auto_stopped

instance.stopped

Jobs

job.queued

job.started

job.finished

Schedules

schedule.failed

schedule.paused

Billing

balance.low

balance.depleted

credits.added

Storage

storage.suspended

storage.reactivated

Webhooks

webhook.test

webhook.disabled

Related