AGENT SKILL

pm-datadog

Datadog connector knowledge and safe action guide.

License 许可
AGPL-3.0
所属仓库
polymetrics-ai/cli
最近更新
2026-09-21
Stars
★ 7
Forks
0

01 前言 Front Matter

name
pm-datadog
description
Datadog connector knowledge and safe action guide.

03 SKILL.md 正文

pm-datadog

Purpose

Reads Datadog monitors, dashboards, dashboard lists, users, SLOs, SLO corrections, scheduled downtimes, notebooks, organizations, hosts, Synthetics tests/locations/variables, and API/application keys, and writes monitor/dashboard/downtime/notebook/SLO/user/event/Synthetics-test/API-key mutations, through the Datadog v1 REST API.

Icon

Capabilities

  • check=true catalog=true read=true write=true query=false
  • Integration type: api

Authentication

  • Use pm credentials add with --from-env or --value-stdin for secret fields.

Configuration

  • base_url
  • api_key (secret) (required)
  • application_key (secret) (required)

ETL Streams

  • monitors:
    • primary key: id
    • cursor: modified
    • fields: created(string), id(integer), message(string), modified(string), name(string), overall_state(string), priority(integer), query(string), type(string)
  • dashboards:
    • primary key: id
    • fields: author_handle(string), created_at(string), description(string), id(string), is_read_only(boolean), layout_type(string), modified_at(string), title(string), url(string)
  • users:
    • primary key: id
    • fields: created_at(string), disabled(boolean), email(string), handle(string), id(string), name(string), status(string), type(string), verified(boolean)
  • slo:
    • primary key: id
    • fields: created_at(integer), description(string), id(string), modified_at(integer), name(string), type(string)
  • downtimes:
    • primary key: id
    • fields: active(boolean), disabled(boolean), end(integer), id(integer), message(string), monitor_id(integer), scope(string), start(integer)
  • dashboard_lists:
    • primary key: id
    • cursor: modified
    • fields: created(string), dashboard_count(integer), id(integer), is_favorite(boolean), modified(string), name(string), type(string)
  • notebooks:
    • primary key: id
    • cursor: modified
    • fields: author_handle(string), created(string), id(integer), modified(string), name(string), type(string)
  • organizations:
    • primary key: public_id
    • fields: created(string), description(string), name(string), public_id(string), trial(boolean)
  • hosts:
    • primary key: id
    • cursor: last_reported_time
    • fields: aliases(array), apps(array), aws_name(string), host_name(string), id(integer), is_muted(boolean), last_reported_time(integer), mute_timeout(integer), name(string), sources(array), up(boolean)
  • slo_corrections:
    • primary key: id
    • cursor: modified_at
    • fields: category(string), created_at(integer), description(string), duration(integer), end(integer), id(string), modified_at(integer), slo_id(string), start(integer), timezone(string), type(string)
  • synthetics_tests:
    • primary key: public_id
    • fields: locations(array), message(string), monitor_id(integer), name(string), public_id(string), status(string), subtype(string), tags(array), type(string)
  • synthetics_locations:
    • primary key: id
    • fields: id(string), name(string)
  • synthetics_variables:
    • primary key: id
    • fields: description(string), id(string), is_fido(boolean), is_totp(boolean), name(string), parse_test_public_id(string), tags(array)
  • api_keys:
    • primary key: key
    • cursor: created
    • fields: created(string), created_by(string), key(string), name(string)
  • application_keys:
    • primary key: hash
    • fields: hash(string), name(string), owner(string)

Sync Modes

  • ETL sync modes: full_refresh_append, full_refresh_overwrite, full_refresh_overwrite_deduped, incremental_append, incremental_append_deduped

Reverse ETL Actions

  • create_monitor:
    • endpoint: POST /api/v1/monitor
    • required fields: name, type, query, message
    • risk: creates a new alerting monitor; low-risk external mutation, no approval required
  • update_monitor:
    • endpoint: PUT /api/v1/monitor/{{ record.id }}
    • required fields: id
    • risk: mutates an existing monitor's alert condition/notification message; a changed query/threshold affects live alerting behavior, approval required
  • delete_monitor:
    • endpoint: DELETE /api/v1/monitor/{{ record.id }}
    • required fields: id
    • risk: irreversibly removes a monitor and its alerting history reference; approval required
  • create_dashboard:
    • endpoint: POST /api/v1/dashboard
    • required fields: title, layout_type, widgets
    • risk: creates a new dashboard; low-risk external mutation, no approval required
  • update_dashboard:
    • endpoint: PUT /api/v1/dashboard/{{ record.id }}
    • required fields: id, title, layout_type, widgets
    • risk: replaces an existing dashboard's full widget layout; external mutation, approval required
  • delete_dashboard:
    • endpoint: DELETE /api/v1/dashboard/{{ record.id }}
    • required fields: id
    • risk: irreversibly removes a dashboard; approval required
  • create_dashboard_list:
    • endpoint: POST /api/v1/dashboard/lists/manual
    • required fields: name
    • risk: creates a new dashboard list (folder); low-risk external mutation, no approval required
  • update_dashboard_list:
    • endpoint: PUT /api/v1/dashboard/lists/manual/{{ record.id }}
    • required fields: id, name
    • risk: renames an existing dashboard list; external mutation, approval required
  • delete_dashboard_list:
    • endpoint: DELETE /api/v1/dashboard/lists/manual/{{ record.id }}
    • required fields: id
    • risk: irreversibly removes a dashboard list (folder); the dashboards themselves are unaffected, approval required
  • create_downtime:
    • endpoint: POST /api/v1/downtime
    • required fields: scope
    • risk: schedules a downtime that silences monitor alerts for the given scope; suppresses real alerting during the window, approval required
  • update_downtime:
    • endpoint: PUT /api/v1/downtime/{{ record.id }}
    • required fields: id
    • risk: mutates an existing downtime's window/scope; changes which alerts are currently suppressed, approval required
  • cancel_downtime:
    • endpoint: DELETE /api/v1/downtime/{{ record.id }}
    • required fields: id
    • risk: cancels a scheduled/active downtime; alerting resumes immediately for its scope, approval required
  • create_notebook:
    • endpoint: POST /api/v1/notebooks
    • required fields: name, cells, time
    • risk: creates a new notebook; low-risk external mutation, no approval required
  • update_notebook:
    • endpoint: PUT /api/v1/notebooks/{{ record.id }}
    • required fields: id, name, cells, time
    • risk: replaces an existing notebook's content; external mutation, approval required
  • delete_notebook:
    • endpoint: DELETE /api/v1/notebooks/{{ record.id }}
    • required fields: id
    • risk: irreversibly removes a notebook; approval required
  • create_slo:
    • endpoint: POST /api/v1/slo
    • required fields: name, type, thresholds
    • risk: creates a new SLO target; low-risk external mutation, no approval required
  • update_slo:
    • endpoint: PUT /api/v1/slo/{{ record.id }}
    • required fields: id
    • risk: mutates an existing SLO's target thresholds; affects SLO burn-rate alerting, approval required
  • delete_slo:
    • endpoint: DELETE /api/v1/slo/{{ record.id }}
    • required fields: id
    • risk: irreversibly removes an SLO and its historical error-budget tracking; approval required
  • create_user:
    • endpoint: POST /api/v1/user
    • required fields: email
    • risk: invites a new user into the Datadog organization with the given role; approval required
  • update_user:
    • endpoint: PUT /api/v1/user/{{ record.handle }}
    • required fields: handle
    • risk: mutates an existing user's role/profile; a changed access_role directly changes that user's permissions, approval required
  • disable_user:
    • endpoint: DELETE /api/v1/user/{{ record.handle }}
    • required fields: handle
    • risk: disables a user's access to the Datadog organization; approval required
  • create_event:
    • endpoint: POST /api/v1/events
    • required fields: title, text
    • risk: posts a custom event into the Datadog event stream; low-risk external mutation, no approval required
  • create_synthetics_api_test:
    • endpoint: POST /api/v1/synthetics/tests/api
    • required fields: name, type, config, locations
    • risk: creates a new Synthetics API test that begins actively probing the configured URL/host on a schedule; low-risk external mutation, no approval required
  • update_synthetics_api_test:
    • endpoint: PUT /api/v1/synthetics/tests/api/{{ record.public_id }}
    • required fields: public_id
    • risk: mutates an existing Synthetics API test's request target/assertions; changes what is actively probed, approval required
  • create_api_key:
    • endpoint: POST /api/v1/api_key
    • required fields: name
    • risk: creates a new organization API key with full agent-submission scope; a newly-minted long-lived credential, approval required
  • update_api_key:
    • endpoint: PUT /api/v1/api_key/{{ record.key }}
    • required fields: key, name
    • risk: renames an existing API key; low-risk external mutation, no approval required
  • delete_api_key:
    • endpoint: DELETE /api/v1/api_key/{{ record.key }}
    • required fields: key
    • risk: irreversibly revokes an organization API key; every agent/integration still using it immediately loses ingest access, approval required

Security

  • read risk: external Datadog API read of monitor, dashboard, SLO, downtime, notebook, organization, host, Synthetics, and API/application key configuration data
  • write risk: external mutation of Datadog monitors, dashboards, downtimes, notebooks, SLOs, users, events, Synthetics API tests, and API keys; create_downtime/update_downtime suppress real alerting for their scope, update_monitor/update_slo change live alerting/burn-rate thresholds, and delete_api_key immediately revokes ingest access for anything still using it, so every write ships an explicit per-action risk string
  • approval: required for every delete_*/cancel_downtime action (irreversible or alerting-suppressing) and for update_monitor/update_downtime/update_slo/update_user/update_synthetics_api_test/create_downtime/create_user (live alerting or access-control side effects); create_monitor/create_dashboard/create_dashboard_list/create_notebook/create_slo/create_event/create_synthetics_api_test/update_dashboard_list/update_api_key are low-risk
  • Never pass secret values in chat, shell arguments, logs, docs, or JSON output.

Commands

Inspect as a manual

pm connectors inspect datadog

Inspect as structured JSON

pm connectors inspect datadog --json

Agent Rules

  • Run pm connectors inspect datadog before creating credentials or plans.
  • Use --json only when the caller needs structured output; use the manual for human-readable guidance.
  • Never ask the user to paste secret values into chat.
  • For reverse ETL writes, create a plan, show the preview, wait for explicit approval, then run with the approval token.

04 同仓库更多 Skills