Better Stack

Connect Better Stack with OAuth so agents can check monitors, open incidents, on-call coverage, and per-source logs

Connect Better Stack to let CloudThinker agents check which monitors are down, read open incidents and who’s on call, and search logs and errors in your telemetry sources. You sign in to Better Stack once through OAuth, so there’s no API token to create or paste.

Prerequisites

• A Better Stack account that can see the monitors, incidents, and telemetry sources you want agents to read.
• Permission in Better Stack to authorize a third-party app.

Setup

1. Start the connection

   Open More → Workspace settings → Connections, select Better Stack, and click Connect. CloudThinker opens Better Stack’s authorization page.

2. Authorize CloudThinker

   Sign in to the Better Stack account agents should use, then approve access. If your browser is signed in to a different Better Stack account, sign out of it first.

3. Return to CloudThinker

   Better Stack sends you back to CloudThinker, which stores the tokens for you.

   Success state: the connection shows Connected. Run the prompt in Verify the connection in a new chat to see what the account can reach.

Connection details

There are no fields to fill in.

Item	Value
Authentication	OAuth. Better Stack issues the tokens and CloudThinker stores them
Endpoint	https://mcp.betterstack.com, Better Stack’s hosted MCP server

Agent requests go from CloudThinker to that server, which reads your Better Stack account. Nothing connects into your network, so there’s no firewall rule to add.

Required permissions

Agents see what the authorizing Better Stack account sees. CloudThinker can’t narrow that grant, so limit access by authorizing with an account that only reaches the teams you want agents to use.

The authorizing account’s own Better Stack permissions are the ceiling. For a read-only connection, authorize with an account that can’t change incidents or dashboards. Choose Manual so a person approves every change, and once CloudThinker turns on Rule enforcement for your workspace, a Deny Rule can block Better Stack changes outright.

Agent capabilities

Capability	What agents can do
Discovery	List monitors, incidents from the last 24 hours, telemetry sources, and status pages in one pass
Uptime	Read monitor availability and response times over a window, plus heartbeat hit rates for cron and scheduled jobs
Incidents and on-call	Read incidents, timelines, and comments, on-call schedules and current rotations, escalation policies, and severities
Status pages	Read status pages, their components, and past reports
Telemetry	Query logs and metrics in a named source, and read dashboards, charts, alerts, errors, applications, and releases
Changes (through your Approvals)	Acknowledge, resolve, reopen, escalate, comment on, or create an incident; publish a status page report or update; create, edit, or delete dashboards, charts, and alerts

Better Stack keys every resource by ID, so agents look up a monitor, incident, or source by name first and confirm the match before changing anything. Acknowledging an incident silences its on-call paging, so approve it only when someone is actually handling the incident. To decide which changes ask first and which are refused outright, see Rules.

Every change an agent makes goes through your workspace’s Approvals first. On the default Auto · Rules + AI · Ask me mode, AI judgment can let a safe change run without asking — and an Allow Rule can too, once CloudThinker turns on Rule enforcement for your workspace. Choose Manual to approve every change yourself.

Better Stack’s MCP server can’t pause or resume a monitor. Set maintenance windows in the Better Stack dashboard.

Verify the connection

Ask CloudThinker in a new chat:

List open Better Stack incidents, monitors that are down, and who is on call now

Expect any open incidents with their monitor and start time, the monitors currently failing, and the current on-call person per schedule. Empty lists are a valid answer when everything is up.

Example prompts

24-hour availability and p95 response time for every production monitor
Count 5xx errors by status code in the api-gateway logs over the last hour
Which heartbeats missed a check-in today?

Better Stack has no global log search: every query runs against one source. Name the source in your prompt, as in api-gateway above, or ask the agent to list sources first.

Troubleshooting

The OAuth flow doesn't complete

Your browser is probably signed in to a different Better Stack account. Sign in to the right account in Better Stack, then click Connect again.

Calls fail with 401 Unauthorized

The OAuth token expired or access was revoked in Better Stack. Open the Better Stack connection, click Connect, and authorize again. The connection keeps its agents, settings, and memory.

A monitor or incident returns 404 Not Found

The resource belongs to a team the authorizing account can’t see, or it was deleted. Ask the agent to list monitors or incidents again to get current IDs.

A log query returns a syntax error

Better Stack log queries use its own query language. The agent loads Better Stack’s query syntax and retries on its own. If it still fails, ask again in plain words and name the source, for example “count 5xx errors in api-gateway over the last hour”.

Requests fail with 429 Too Many Requests

You hit Better Stack’s rate limit. Agents back off and retry. If it keeps happening, narrow the time window or the number of monitors in your question.

Security

• Least privilege — grant only the permissions the agents need for your use case; start read-only and widen later.
• Read-only by default — use read-only credentials unless you want agents to make changes through this connection.
• Rotate credentials — rotate keys and tokens on your normal schedule; CloudThinker picks up the new value when you update the connection.
• Revoke on offboarding — remove the credential at the provider when you delete a connection or a teammate leaves.

• Revoke when unused: remove CloudThinker’s authorization in Better Stack if you stop using the connection, then delete the connection in CloudThinker.
• Switching accounts: disconnect, then connect again while signed in to the other Better Stack account.

Related

Approvals

How CloudThinker asks before an incident or dashboard change runs.

PagerDuty

Incident alerting and on-call schedules.