ingestlayer/recipes

Monitor cron-job health in Slack

Hear about the scheduled jobs that fail or run long — and only those — so a silent backup or a stalled sync becomes a message, not a surprise.

01source

sourcesdk.eventTypeScript SDK
matchcron.finished

02pipeline · 1 steps

  • 01CTLfilter.matchstatus = failed or ran long

03destinations · 1

  • toslackSlack
    channel#alerts

the event

You emit cron.finished with this shape. The TypeScript SDK keeps the call type-safe, and the event is stored whole — so every field below is available to the pipeline by name.

  • jobstringjob name
  • statusstringok | failed
  • duration_msnumber
  • expected_msnumbertypical runtime
  • exit_codenumber

emit it

From your code with the TypeScript SDK — or any language over the REST endpoint and signed webhook ingress.

emit cron.finished
import { ingest } from "@ingestlayer/sdk";

await ingest("cron.finished", {
  job:         "nightly-backup",
  status:      ok ? "ok" : "failed",
  duration_ms: Date.now() - startedAt,
  expected_ms: 120000,
  exit_code:   code,
});

route it to Slack

Post to any channel in your workspace. Connect once with OAuth, pick the channel per pipeline.

  1. 01

    connect your workspace

    Authorize the ingestlayer Slack app over OAuth from the destinations page. We hold only a channel-scoped bot token, in-region, in the same KMS as your other credentials.

  2. 02

    pick a channel

    Choose any public channel, or invite the bot to a private one. The channel is set per pipeline, so different events can land in different places.

  3. 03

    map the message

    Reference event fields with $event.* in the message template. The default renders a titled block with the event name and its key fields.

in slackdelivered
┌─ #alerts ──────────────────────────────┐
│  ingestlayer  APP                       │
│  user.signed_up                         │
│  email   ada@acme.com                   │
│  plan    pro                            │
│  source  marketing-site                 │
└─────────────────────────────────────────┘

notes

questions

Can I only hear about problems?
Filter to failures and jobs that overran their expected_ms; healthy runs pass silently to the warehouse for the record.
How do I catch a job that never ran?
Pair this with a scheduled check that emits when an expected cron.finished is missing — absence is the signal there.
Can different jobs alert different teams?
Branch on the job name so a failed billing run reaches finance and a failed deploy reaches engineering.
build this pipelineor read the quickstart →

cron-job health, routed elsewhere

more, into Slack