Triage alerting Datadog monitors by service
For Datadog alert triage, Manus lists the monitors in Alert or Warn, groups them by service and puts the one that fired first at the top, so you know where to start.
Find out what broke in Datadog without digging through dashboards
Connect Datadog and ask in plain words, like why payments got slow after 11 PM. Manus checks alerting monitors, uses AI log analysis to sort through the logs, reads metrics and traces, compares before and after the last deploy, then writes an incident summary with the likely cause. It looks things up and reports back. With the default Datadog tools, it can’t roll back deploys, restart services or change your monitors.
First, connect Datadog in Manus
Incident timelineservice:payments-api env:prodIllustration
- 23:12Deploypayments-api v2.4.1 deployedDeployment event from your pipeline
- 23:16AlertMonitor: p95 latency above 800 msStatus changed from OK to Alert
- 23:19Errors5xx errors rise to 6.8%Up from 0.3% before the deploy
- 00:04RecoveredLatency back under 200 msAfter the team redeployed v2.4.0 at 23:58
Datadog AI log analysis and alert triage
Ask what’s firing and what the logs say. Manus reads your alerting monitors and searches the logs, then tells you where to look first. It works as Datadog AI monitoring help for whoever is on call.
Search Datadog logs and group errors into patterns
Name a service and a time window. Manus runs a Datadog log search for errors and warnings, counts them by message and shows a few real lines from each group.
Why is my service slow? Check the trace and the last deploy
Manus finds the span that takes the longest and lines up latency and errors against your deployment events.
Find the slow step in a Datadog trace
Manus pulls the slowest requests for a service, opens one trace and points to the span that takes the most time.
Compare errors before and after a deployment
Use Manus as your Datadog AI SRE helper for the first check after a release: it finds the deployment events and looks for errors after deployment, comparing the error rate and p95 latency in the hour before and after each one.
Write a Datadog incident summary and postmortem draft
Manus reads the incident and the alerts and deploys around it, so you start from a timeline instead of a blank page.
Overnight Datadog summary
Since 18:00 yesterday: 4 monitors alerted and 1 incident opened. Everything is back to OK.
Turn an incident into a summary and a postmortem draft
Manus reads the incident, builds the timeline from the alerts and deployment events around it, then writes an incident summary and an incident postmortem draft for you to edit.
Get last night’s alerts in one message every morning
A scheduled task checks Datadog at 9 AM on weekdays and sends you what alerted and which incidents opened overnight.
See how scheduled tasks workMap service dependencies and Datadog cloud cost changes
Before you escalate, see what calls the service and who owns it. With the right permission, compare cloud costs month over month too.
Check who depends on a service before you escalate
Manus looks the service up in the Datadog catalog and lists what calls it, what it calls and which team owns each one.
Compare cloud cost changes month over month
If your team uses Datadog Cloud Cost Management and your role can read it, Manus compares two months and shows which services changed the most.
Datadog MCP server setup: permissions and tools
For Datadog MCP setup, there is nothing to install. Datadog is already in Manus Connectors: connect your account once, and any task can use the Datadog MCP tools your role allows.
Open Connectors
In Manus, open Connectors, choose Datadog, select Connect and connect your Datadog account.
Check your Datadog access
MCP access must be on for your organization, and your role needs MCP Read. The Datadog Standard Role has it by default.
Ask in any task
Ask about alerts, logs, traces or incidents in your own words. Manus sees only what your Datadog role can see.
More Datadog integrations with Manus
Questions about the Datadog MCP connector
How do I connect Datadog to Manus?
Open Connectors in Manus, choose Datadog, select Connect and connect your Datadog account. MCP access must be turned on for your Datadog organization, which an admin manages in Organization Settings, and your role needs the MCP Read permission. The Datadog Standard Role includes it.
Will Manus change my production setup or alerts?
Not with the default tools. The default Datadog tools read logs, metrics, traces, monitors and incidents. The only things they can create or change are Datadog notebooks and dashboards, and only when you ask. They have no tools to roll back deploys, restart services or edit monitors. Optional toolsets can add write tools, such as creating monitors, so set up only the ones you need. An admin can also turn off MCP write access for everyone in Organization Settings.
What Datadog data can Manus see?
Only what your own Datadog account can see. The Datadog MCP server uses your access, so your roles and data access rules still apply, and each kind of data needs its usual Datadog permission, such as Logs Read Data for logs or APM Read for traces.
Which Datadog MCP tools and toolsets does Manus use?
It starts with the core Datadog MCP server tools for logs, metrics, traces, monitors, incidents, events, dashboards, hosts, services and notebooks. The connector can also be set up with optional Datadog MCP toolsets for specific Datadog products, such as APM or Database Monitoring.
Which Datadog sites are supported?
The Datadog MCP server isn’t available on Datadog for Government sites. For the other Datadog sites, check the Datadog MCP server documentation for the current list.
Is Manus the same as Datadog Bits AI?
No. Bits AI is Datadog’s own AI teammate, built into Datadog and billed by Datadog. Manus is a separate AI agent that reads your Datadog data through the Datadog MCP server and can carry the results into your other apps, documents and scheduled tasks.
I use Datadog in Claude or ChatGPT. Can I use it in Manus too?
Yes. The Datadog Claude connector and the Datadog ChatGPT integration, which is in Preview for US1 only, both run on the Datadog MCP server, and the Datadog connector in Manus uses that same server. Connect it once in Manus Connectors, and any task can use it.
Can Manus post the incident summary to Slack or open a ticket?
Yes, if those apps are connected too. Manus can post the summary to a Slack channel and open a follow-up issue in Linear, GitHub or Jira. Ask it to show you the text first and post only after you say OK.
Can Manus check Datadog on a schedule?
Yes. Create a scheduled task, for example every weekday at 9 AM, and Manus checks the alerts and incidents from the night before and sends you a short summary grouped by service.
Can my team see what Manus looked up in Datadog?
Yes. Datadog records every MCP server tool call in Audit Trail, including the tool, what it asked for, which user it ran as and which client sent it.
How does Datadog MCP pricing work?
Any Datadog charges are part of your company’s Datadog account and plan. For current Datadog MCP server pricing and usage limits, check the Datadog documentation.