How to Monitor Your Automations with Sentry and DataDog

An automation that fails silently for 3 days is worse than no automation. Here's the monitoring stack we put on every Nexora-built workflow.

The 3 layers of monitoring

1. Error tracking (Sentry) — catches exceptions and groups them
2. Metrics + logs (DataDog or Grafana Cloud) — tracks run frequency, latency, success rate
3. Heartbeats (Better Stack or Healthchecks.io) — alerts if a cron didn't fire

Sentry: 10-minute setup

import sentry_sdk
sentry_sdk.init(dsn="https://abc@sentry.io/123", traces_sample_rate=0.1)

try: run_automation() except Exception as e: sentry_sdk.capture_exception(e) raise ```

Sentry's free tier covers 5K events/month — enough for most automation workloads.

DataDog: real-time metrics

from datadog import statsd
statsd.increment("nexora.automation.run", tags=["job:invoice_sync"])
statsd.timing("nexora.automation.duration", elapsed_ms, tags=["job:invoice_sync"])

DataDog is expensive ($31/host/month). For cost-conscious teams, Grafana Cloud free tier offers similar features.

Heartbeats: the most underrated layer

Healthchecks.io (free) gives you a unique URL per cron. Your automation pings it after every successful run. If the ping doesn't arrive within the expected interval → email/Slack alert.

curl https://hc-ping.com/your-uuid > /dev/null 2>&1

Add this as the last line of every cron job. Done.

What we monitor on every Nexora gig

• Run counts (per-day, per-week trend)
• Success rate (% non-error runs)
• Duration P50, P95, P99
• Last successful run timestamp
• Error rate by error type

Alert thresholds

• Success rate < 95% over 1 hour → page
• No run in 2x expected interval → page
• P95 duration > 3x normal → notify

Hire a monitoring + DevOps expert →