Agent Skills › latitude-dev/latitude-llm › managing-maintenance-windows

managing-maintenance-windows

GitHub

管理 Latitude 生产环境维护窗口，通过 Pulumi 切换 ALB 配置将流量重定向至状态页。支持启用、禁用及验证维护模式，包含安全检查清单和具体操作命令。

.agents/skills/managing-maintenance-windows/SKILL.md latitude-dev/latitude-llm

Trigger Scenarios

启用或停止生产维护模式验证维护模式是否生效准备维护窗口

Install

npx skills add latitude-dev/latitude-llm --skill managing-maintenance-windows -g -y

More Options

Non-standard path

npx skills add https://github.com/latitude-dev/latitude-llm/tree/development/.agents/skills/managing-maintenance-windows -g -y

Use without installing

npx skills use latitude-dev/latitude-llm@managing-maintenance-windows

指定 Agent (Claude Code)

npx skills add latitude-dev/latitude-llm --skill managing-maintenance-windows -a claude-code -g -y

安装 repo 全部 skill

npx skills add latitude-dev/latitude-llm --all -g -y

预览 repo 内 skill

npx skills add latitude-dev/latitude-llm --list

SKILL.md

Frontmatter

{
    "name": "managing-maintenance-windows",
    "description": "Enables or disables Latitude production maintenance mode by redirecting all publicly exposed production services to the Better Stack status page. Use when asked to start, stop, toggle, verify, or prepare a maintenance window."
}

Managing maintenance windows

Use this skill when a user asks to enable, disable, verify, or prepare Latitude's production maintenance window.

What maintenance mode does

Maintenance mode is an ALB-level switch in infra/lib/alb.ts.
When enabled, all public production service hostnames redirect to https://status.latitude.so/: console.latitude.so, api.latitude.so, ingest.latitude.so, and bull-board.latitude.so.
The switch is controlled by the Pulumi production config key enableWebMaintenanceRedirect.
Do not implement this in apps/web: if a service is unhealthy, app-level redirects may not run.

Enable maintenance mode

From infra/:

pulumi config set enableWebMaintenanceRedirect true --stack production
pulumi preview --stack production
pulumi up --stack production

Then verify from outside the VPC:

curl -I https://console.latitude.so/
curl -I https://api.latitude.so/health
curl -I https://ingest.latitude.so/health
curl -I https://bull-board.latitude.so/

Expected result: all requests return an HTTP redirect with Location: https://status.latitude.so/.

Disable maintenance mode

From infra/:

pulumi config rm enableWebMaintenanceRedirect --stack production
pulumi preview --stack production
pulumi up --stack production

Then verify:

curl -I https://console.latitude.so/
curl -I https://api.latitude.so/health
curl -I https://ingest.latitude.so/health
curl -I https://bull-board.latitude.so/

Expected result: none of the requests redirects to https://status.latitude.so/. Services may return normal responses or auth redirects depending on the requested path.

Safety checklist

Before enabling:

Confirm the user wants all publicly exposed production services redirected.
Confirm https://status.latitude.so/ is live and has the relevant Better Stack maintenance or incident notice.
Run pulumi preview --stack production and check that the ALB HTTPS listener default action and all public host rules change from forwarding to redirecting.
Apply only after the user has approved the production change, unless they already explicitly asked to execute it.

Before disabling:

Confirm the maintenance window is complete or the user explicitly asked to restore service.
Run pulumi preview --stack production and check that the ALB HTTPS listener default action and all public host rules change from redirecting back to forwarding.
Verify every public hostname after pulumi up.

Notes for agents

Use eval "$(mise env)" if local shell commands cannot find node, pnpm, or Pulumi-related tooling.
Do not change the hardcoded redirect target to an incident-specific URL. The reusable target is the status page root, https://status.latitude.so/.
If Pulumi credentials or stack access are unavailable, report the exact command the operator should run and do not fake success.