# Watch signals on a schedule and report anomalies

> Watch signals on a schedule and report anomalies: 11 companies fully do it, 1 with a limit, 20 carry the tag. Every listing in Observability & incidents with the reason and a source.

- Canonical: https://www.anchorterminal.com/niches/observability.scheduled-watch/
- Markdown: https://www.anchorterminal.com/niches/observability.scheduled-watch/index.md (~1,250 tokens)
- Slim: https://www.anchorterminal.com/niches/observability.scheduled-watch/index.min.md (~330 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/niches/observability.scheduled-watch/index.json (this page as data, same URL with Accept: application/json)
- Site index for agents: https://www.anchorterminal.com/llms.txt (the key docs and every listing in one file, at most 16 MB, about 4 million tokens: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- MCP: https://www.anchorterminal.com/mcp (the whole catalogue as tools, no key)
- DNS: dig +short TXT <slug>.grade.anchorterminal.com (the grade as a DNS TXT record, for a sandbox with no HTTP where names still resolve; also top., best.<category>. and <a>.vs.<b>. under the zone, and the same under grade.letme.dev)
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)
- Updated: 2026-11-15

Four or more companies' docs show a full fit.

| Companies, full fit | With a limit | Carry the tag | Share of the category's work | First-hand accounts |
| --- | --- | --- | --- | --- |
| 11 | 1 | 20 | 14 | 7 |

## What counts as done

The run completes unattended with a non-interactive credential, runs the fixed queries without an extra LLM call per query, and sends a digest that lists only new or changed findings with links.

How agents ask for it: send me a daily digest of new errors; check the logs every morning; tell me if anything looks off overnight; run these queries every day

## Does the job

The docs show it meets the test above, and nothing we read says otherwise.

- **Axiom MCP Server (Axiom, Inc.)**: Header auth with a personal access token whose expiry is optional runs without a browser; fixed APL queries run as given with no model call; cost limits cap spend per hour and day.
- **Better Stack (Better Stack, Inc.)**: A team API token in the Authorization header runs without a browser, and fixed SQL goes straight to Execute query with no model call.
- **Causely MCP server (Causely, Inc.)**: A personal or tenant-wide token in X-Causely-Client-Basic skips the browser; tools take structured arguments, so fixed calls need no model call, and team_health or get_environment_health give a digest.
- **Cronitor MCP server (Cronitor)**: The long-lived SDK Integration key works as a Bearer header with no browser; get_status gives an account rollup with failing monitors and list_issues a summary mode, with no model call per query.
- **Datadog MCP Server (Datadog, Inc.)**: API and application keys sent as headers run without a browser, queries are passed as written, and the burst limit is 50 calls per 10 seconds.
- **Grafana MCP Server (Raintank Inc., d/b/a Grafana Labs)**: A service account token in GRAFANA_SERVICE_ACCOUNT_TOKEN runs unattended, and fixed PromQL or LogQL run as written with no model call.
- **Honeycomb MCP (Hound Technology, Inc. d/b/a Honeycomb)**: A Management API key sent as Authorization: Bearer KEY_ID:SECRET_KEY serves unattended agents, and run_query takes a fixed spec with no model call; run_query allows 100 calls a minute.
- **incident.io API (Pineapple Technology Ltd.)**: API keys authenticate as a service actor and don't expire until deleted; incident_list, alert_list and incident_stats take structured filters, so a fixed digest needs no model call per query.
- **New Relic MCP Server (New Relic, Inc.)**: A User API key works for automated scripts without a browser, and execute_nrql_query runs fixed NRQL with no model call; the AI analysis tools are OAuth only.
- **PagerDuty MCP Server (PagerDuty, Inc.)**: A User API key in Authorization: Token token= or an app OAuth token from client credentials runs unattended, and browse calls take structured filters with no model call.
- **Rootly MCP Server (Rootly Inc.)**: A Global, Team or Personal API token as a Bearer header runs unattended, and list tools take structured filters with no model call.

## Does the job, with a limit

It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.

- **Sentry MCP (Functional Software, Inc. d/b/a Sentry)**: A Sentry-Bearer token header now works on the hosted server, and acc-01 runs a daily digest, but search tools add a model call per query and the maintainers declined a raw query field (#471).

## What stops agents finishing

- browser sign-in that expires and needs a person
- LLM token cost of reading raw high-volume streams
- natural-language-only query tools that add a model call per query

## Around it

Before: webhooks.schedule-call. After: messaging.post-summary, email.send, observability.investigate-alert.

## Where this comes from

Weight 14 of 100 in Observability & incidents: 7 of 42 accounts, mostly from the Agentic Monitoring thread; that thread also argues against agents reading raw streams (acc-27)

As data: https://www.anchorterminal.com/api/v1/use-cases/observability.scheduled-watch.json, or the MCP tool get_use_case with id observability.scheduled-watch. Draft, read 2026-11-15 from public docs and first-hand accounts, not yet reviewed by a person.

---

Query this catalogue directly over MCP: https://www.anchorterminal.com/mcp (Streamable HTTP, no key). How: https://www.anchorterminal.com/agents/#mcp
