What you get with and without the agent
DigitalOcean describes Monitoring as a free, opt-in service. Without the agent, Droplet graphs show CPU usage, public and private bandwidth and disk I/O. Installing the metrics agent adds load averages, memory usage and disk usage, and resource alerts only work on Droplets or tags with the agent installed. DigitalOcean's limits page notes that enabling the agent is one way: you cannot revert to basic Droplet graphs afterwards.
To enable it at creation, tick Improved Metrics and monitoring (Free) under Monitoring, or pass monitoring true in the API or --enable-monitoring in doctl. On an existing Droplet, the documented install is a one-line curl command that pipes DigitalOcean's agent install script to sudo bash (copy it from the install-agent page in the sources), with a review-first option of downloading the script to /tmp and reading it before running. Check it with systemctl status do-agent.
| Metric | Without agent | With agent | Alertable (agent required) |
|---|---|---|---|
| CPU utilisation | Yes | Yes | Yes |
| Public and private bandwidth, in and out | Yes | Yes | Yes |
| Disk read and write I/O | Yes | Yes | Yes |
| Memory utilisation | No | Yes | Yes |
| Disk utilisation | No | Yes | Yes |
| 1, 5 and 15 minute load average | No | Yes | Yes |
How alert policies work
DigitalOcean's alert documentation lists 12 Droplet metrics: three load averages, memory, disk and CPU utilisation percent, disk read and write I/O, and public and private inbound and outbound bandwidth. In the control panel the rule is is above or is below, the window is 5, 10 or 30 minutes or 1 hour, and the default threshold is 70. A policy can target all Droplets, one Droplet or a tag. You must pick email, Slack or both. Triggered alerts appear on the Resource Alerts tab, and a second notification arrives when the condition clears.
Two constraints affect who gets paged. Email alerts go to the account's default address or verified members of the same team, and the limits page says you cannot set a separate alert email address for an individual account. DigitalOcean suggests a team email or a shared Slack channel when several people need the alert.
Starting thresholds for an LMS Droplet
These are editorial starting points, not DigitalOcean guidance. Watch a normal week of graphs and adjust so that each alert means someone should act.
| Alert | Rule | Window | Why this level |
|---|---|---|---|
| CPU utilisation | Above 80% | 10 min | Sustained CPU above this during a live class means PHP requests are queuing; short spikes during imports are normal |
| Memory utilisation | Above 85% | 10 min | Leaves headroom before the kernel kills MySQL or PHP-FPM workers |
| Disk utilisation (warning) | Above 80% | 5 min | Time to clear logs, old backups or move media off the Droplet |
| Disk utilisation (urgent) | Above 90% | 5 min | A full disk stops MySQL writes, which breaks logins, progress saving and checkout |
| 5-minute load average | Above the vCPU count | 10 min | More runnable work than cores for a sustained period |
| Public outbound bandwidth | Above about 3x your normal peak | 30 min | Catches hotlinked downloads or abuse before a transfer bill arrives |
Add Uptime checks for what learners see
Resource alerts tell you the server is strained, not that learners can log in. DigitalOcean Uptime checks the health of a URL or IP address and can alert on downtime, high latency and SSL certificates about to expire, by email or Slack. Its pricing page says each check costs $1.00 per month, billed hourly, alerts cost nothing extra, and each invoice gets one free check a month, shown as 672 hours of credit, which is spread across checks when no single check runs for the whole month.
Worked example: checks on the homepage, the login page and the checkout page cost 3 x $1.00 minus the one free check, or $2.00 a month. Monitoring itself adds nothing. For cron heartbeats and checks on specific page content, see our membership uptime monitoring guide.
Who gets paged for what
| Alert | Channel | Who | First action |
|---|---|---|---|
| Uptime: login or checkout down | Email and Slack | Site owner and whoever holds server access | Check the Droplet graphs, then PHP-FPM and MySQL status |
| Uptime: SSL expiring | Whoever manages certificates and DNS | Check the renewal timer or certificate source | |
| Disk above 90% | Email and Slack | Server administrator | Find large files under logs, backups and uploads |
| Memory or CPU sustained high | Slack | Server administrator | Check for a stuck cron job, a traffic spike or a plugin update; consider resizing |
| Disk above 80%, bandwidth unusual | Slack only | Server administrator during working hours | Plan cleanup or move downloads to object storage |
Setup checklist
- Confirm do-agent is running on every production Droplet; without it, resource alerts cannot target the Droplet.
- Tag production Droplets (for example lms-prod) and create alert policies by tag so a rebuilt or resized Droplet stays covered.
- Create the six policies above and send each to the right channel.
- Add Uptime checks for login and checkout, not just the homepage.
- Trigger one alert on purpose (for example a short CPU stress test on staging) to confirm the email or Slack message arrives and the cleared notice follows.
- Review thresholds after the first busy cohort.
Questions before you decide
Is DigitalOcean Monitoring free?
Yes. DigitalOcean describes Monitoring as a free, opt-in service. Uptime checks are billed separately at $1.00 per check per month after one free check.
Why can I not select my Droplet when creating an alert?
DigitalOcean's alert documentation says only Droplets with the metrics agent installed are available to select. Install do-agent first.
Which notification channels do alerts support?
Email, Slack or both. Email goes to the account default address or verified team members; you must choose at least one method.
Can I go back to basic graphs after installing the agent?
No. DigitalOcean's monitoring limits page says that after you enable the metrics agent you cannot revert to basic Droplet graphs.
Sources & verification
Product details and prices can change. Check the linked provider before buying.
- DigitalOcean Docs: Monitoring Accessed 30 Oct 2026
- DigitalOcean Docs: Monitoring features Accessed 30 Oct 2026
- DigitalOcean Docs: Monitoring limits Accessed 30 Oct 2026
- DigitalOcean Docs: Set up resource alerts Accessed 30 Oct 2026
- DigitalOcean Docs: Install the metrics agent Accessed 30 Oct 2026
- DigitalOcean Docs: Uptime Accessed 30 Oct 2026
- DigitalOcean Docs: Uptime pricing Accessed 30 Oct 2026
Sources link directly to providers. Product buttons may use separately labeled affiliate links. Read our disclosure.