Skip to main content

Daily Datadog Health Digest

Daily session that scans Datadog for errors and posts a health digest to Slack.
AuthorCognition
CategoryAutomations
FeaturesSchedules, MCP, Integrations
1

Enable the Datadog MCP

Go to Settings > MCP Marketplace and search for Datadog. Click Enable and add two secrets:If your Datadog instance uses a custom site (e.g., datadoghq.eu), also set the optional DATADOG_SITE variable.
2

Create the daily schedule

Navigate to Schedules in the left sidebar of app.devin.ai and click Create schedule:
  • Frequency: Daily — pick a time 30–60 minutes before standup so the team can review findings during the meeting
  • Slack channel: Select a channel (e.g., #ops-alerts) so your team gets the digest automatically
  • Prompt: Be specific about thresholds — vague instructions like “look for issues” produce noisy digests that get ignored
Add Knowledge entries about your services so Devin can distinguish normal from abnormal — for example: “The payments service normally runs at 200ms p99; anything above 400ms is a concern. The search service has a known baseline error rate of 0.5%.”
3

What your team sees in Slack

Each morning, Devin queries Datadog via MCP, checks results against your thresholds, and posts a digest to your channel:
Daily Health Digest — Feb 10, 2026

CRITICAL
  payments-service: Error rate at 3.2% (threshold: 1%)
  Started 4h ago, correlates with deploy #487.
  -> Check /api/checkout handler and recent changes.

WARNING
  search-service: p99 latency at 620ms (threshold: 500ms)
  Gradual increase over 3 days. Likely index degradation.
  -> Review query plans for the product search endpoint.

  web-app: Memory at 82% (threshold: 80%)
  Trending upward since Tuesday.
  -> Investigate session handler for possible memory leak.

INFO
  All other services within normal parameters.
  2 monitors recovered in past 24h: cdn-latency, db-replication-lag.
Reply in the Slack thread to dig into any finding — Devin keeps full context from the scan:
4

Schedule weekly knowledge updates

Your daily health checks generate a stream of knowledge suggestions — especially when you reply to findings with feedback like “this is a known baseline” or “ignore this monitor during deploys.” Schedule a weekly Advanced Devin session to process those suggestions so future runs get smarter automatically.Set this to run weekly (e.g., Monday morning) as an Advanced Devin session so it has access to the Knowledge management tools. Over time, this feedback loop means your daily digests become more accurate — fewer false alarms, better severity ratings, and smarter recommendations.