Schedules

Schedules provide a way to run Speedscale actions on a regular basis. Actions like creating a fresh snapshot from a running service or running a replay. Running replays on a schedule enables regular testing when running Speedscale inside CI/CD is not feasible.

Cron Expressions

Speedscale schedules use a cron expression to define when the actions should run, but there are several predefined intervals to get up and running right away.

The cron expression here, 0 15 * * 1-5, defines a schedule which will run every day at 15:00 UTC, Monday through Friday.

Visit crontab.guru for help creating and understanding cron expressions.

Actions

There are several actions to choose from which can be re-ordered as necessary.

Snapshot

Creates a fresh snapshot from a service running in your cluster. The service must have recorded traffic during the window specified or the snapshot will fail.

Replay

Runs a replay, generating a report, from an existing snapshot. The replay may use an existing snapshot, or one created from a previous action.

Sidecar

Ensures the Speedscale sidecar exists, or does not exist, on a workload running in your cluster. A workload must have the sidecar attached to capture traffic. This action is useful for ensuring a workload is capturing traffic, or turning capture on and off to capture for a schedule without capturing traffic continuously.

info

The sidecar action asynchronously sends a command to add the sidecar to a workload but does not wait for it to complete. Traffic will also take up to 2 minutes to be visible in the Speedscale cloud after capture so it may be necessary to use a Wait action after the sidecar action to achieve consistent results.

Wait

Waits before advancing to the next action. This is useful for waiting for traffic after ensuring a sidecar exists on a workload, or when some out of band work is performed in between actions.

Notify

Sends a notification when the action is reached during execution. Place this action anywhere in the sequence — for example, at the end to report the overall outcome, or immediately after a replay to alert on test failures.

Trigger conditions

Trigger	When it fires
Always (default)	Every time the action is reached, regardless of prior results
On failure	Only when at least one earlier action in the job recorded an error
On success	Only when all earlier actions completed without error

Channels

Webhook sends an HTTP POST to a URL you provide. By default the request body is a structured JSON object:

{
  "job_id": "my-job",
  "job_description": "Nightly regression",
  "execution_id": "3f7a...",
  "start_time": "2024-05-01T15:00:00Z",
  "status": "failure",
  "failed_tasks": [
    { "index": 1, "message": "Error: report completed with errors: [timeout]" }
  ],
  "snapshot_id": "abc123",
  "report_id": "def456",
  "schedule_url": "https://app.speedscale.com/schedules/my-job",
  "report_url": "https://app.speedscale.com/reports/def456"
}

To pass additional headers (for example, an authorization token) use the Headers field when configuring the action.

Message template

Some services — including Slack — require a specific JSON shape and will reject the default payload. Set the Message template field to a Go template string; when present the request body becomes {"text": "<rendered>"} instead.

Available template variables:

Variable	Description
{{.JobID}}	Schedule ID
{{.JobDescription}}	Schedule description
{{.ExecutionID}}	Unique ID for this run
{{.StartTime}}	Execution start time (UTC)
{{.Status}}	success or failure
{{.FailedTasks}}	Slice of failed actions; each has .Index and .Message
{{.SnapshotID}}	ID of the most recent snapshot created in this run
{{.ReportID}}	ID of the most recent replay report created in this run
{{.ScheduleURL}}	Link to this schedule in the Speedscale dashboard
{{.ReportURL}}	Link to the most recent replay report in the dashboard

Slack

1. Create a Slack incoming webhook for the channel you want to post to and copy the webhook URL.

2. Paste the URL into the URL field.

3. Set Message template to a string such as:

   *{{.JobID}}* finished with status *{{.Status}}*
   Started: {{.StartTime}}
   Schedule: {{.ScheduleURL}}{{if .ReportURL}}
   Report: {{.ReportURL}}{{end}}

   The dashboard's Use Slack default button pre-fills this template for you.

   For failure details you can iterate over FailedTasks:

   *{{.JobID}}* failed — {{range .FailedTasks}}task {{.Index}}: {{.Message}} {{end}}

Slack validates that the posted JSON contains a text field and returns 400 no_text without one, so the Message template field is required when posting to Slack.

Email support is coming in a future release.

tip

Combine the Notify action with On failure and continue_on_failure enabled to get an alert any time a scheduled replay misses its goals without stopping the rest of the job.