Open source · Apache 2.0 · Self-hosted

Know when async processes miss their SLO.
And why.

Account activations, payments, KYC checks — when they get stuck, customers notice first. ToTrackIt tracks every run against its deadline, alerts through the stack you already page on, and shows which country, channel, or provider the problem is hiding in.

View on GitHub docker compose up -d

The pager stays yours — Datadog, Prometheus, Grafana. ToTrackIt is where the alert lands.

account-activation SLO 1h
act-9f27 country:DE kyc:provider-x
ON TRACK MISSED
Impacted tags · last 1h
country:DE
14 overdue
kyc:provider-x
9 late
country:FR
0 problems
p90 latency — DE 78m · FR 9m

How it works

Two API calls in. Your alerting out.

1

Instrument the process

Start each run with a deadline and business tags. Complete it when it finishes. That's the whole integration.

POST /processes/account-activation
{ "id": "act-9f27",
  "deadline": now + 1h,
  "tags": [country:DE, kyc:provider-x] }

PUT /processes/…/act-9f27/complete
2

Alert where you already page

ToTrackIt exposes deadline-shaped Prometheus metrics. Point a Datadog monitor or Prometheus alert at them — no new pager, no migration.

processes_overdue_current deadline_missed_total completed_on_time_total completed_late_total
3

Diagnose in one glance

The alert's deep link lands on the process page: every run in the period, impacted tags, and completion latency by segment — avg, p50, p90, p99.

# from the alert payload
"url": "https://totrackit.you.com/?name=account-activation"

Observability-native

Speaks fluent Prometheus

Scrape /prometheus with the Datadog Agent or Prometheus. Build a metric-based SLO — "99% of activations complete within 1 hour" — with error budgets and burn-rate alerts in the tool your team already trusts.

  • Per-process overdue gauge, refreshed every 30s
  • On-time / late / missed counters for SLO math
  • Webhook with deep link when a deadline is breached

What's in the box

Small surface. Sharp answers.

SLO deadlines per run

On-track, missed, on-time, late — computed in real time from timestamps, never stale.

Root cause by tags

See instantly that the stuck runs share a segment: country, channel, provider, customer tier.

Latency by segment

avg / p50 / p90 / p99 completion time per tag. Slow segments stand out before they breach.

Per-process pages

A shareable URL per process with a period picker, impact breakdown, and every run.

Webhooks that land on the answer

Deadline-missed notifications carry tags, context, and a deep link to the affected process.

Self-hosted, Apache 2.0

One Docker Compose file: API, UI, PostgreSQL. Java 21 + Micronaut. Your data stays yours.

Track your first process in five minutes.