Daily DigitalOcean infrastructure health digest in Slack

Every weekday at 8:00am, post a DigitalOcean infrastructure health digest to a Slack channel. The goal is one consolidated morning read across Droplets, managed databases, and App Platform, so the team does not have to piece together fragmented per-metric monitoring alerts.

Keep the run settings at the top so they are easy to change: the schedule (default every weekday at 8:00am), the target Slack channel (for example #infra-health), and the thresholds for what counts as running hot. Default thresholds: CPU averaging above 85% over the last hour, memory usage above 90%, and disk usage above 85% (less than 15% free). Treat all of these as configurable inputs.

First check the Droplets. Use List All Droplets to get every Droplet, then for each one use Retrieve Droplet Metrics to pull recent CPU, memory, and disk figures. Flag any Droplet that is over the CPU, memory, or disk threshold, and note which metric it is and the value.

Next check managed databases. Use List All Database Clusters and read each cluster's status. Flag any cluster that is not in a healthy, online state, for example one that is creating, migrating, resizing, or reporting a failed or degraded status.

Then check App Platform. Use List All Apps to get every app, and for each app use List App Deployments to find its most recent deployment. Use Retrieve an App Deployment when you need the detailed phase or progress for that latest deployment. Flag any app whose most recent deployment did not succeed, for example one in a failed, error, or canceled phase.

Finally, have the agent summarize overall health into a short morning digest and post it with the Slack Send a Message action to the configured channel. Lead with anything unhealthy first, grouped and specific: a Droplet running hot on CPU or low on disk, a database cluster not in a healthy state, or an app whose most recent deployment failed. Name the resource and the reason for each problem. After the problems, add a short green all-clear line that summarizes everything that looks fine, for example how many Droplets are within thresholds, how many clusters are online, and how many apps are on a healthy latest deployment. If nothing is unhealthy, lead with the all-clear line. Format the message with Slack mrkdwn so it is quick to scan, and keep it concise.