Manage incidents from the CLI

Use @siteqwality/cli version 0.1.2 or later with Node.js 22 or later. Incident commands use the account API and do not need a RUM application ID.

npm install --save-dev @siteqwality/cli@0.1.4
export SITEQWALITY_API_KEY='your-incident-api-key'
npx siteqwality incidents list --status active --limit 50 --offset 0
npx siteqwality incidents ack "$INCIDENT_ID" --by "$MEMBER_ID"
npx siteqwality incidents resolve "$INCIDENT_ID" --message 'Service recovered'
npx siteqwality incidents update "$INCIDENT_ID" --title 'Checkout unavailable' --severity critical

Keep the API key in your shell or CI secret store. Listing requires read:incidents; changes require write:incidents. The key is never a command-line flag. JSON results go to stdout, and errors return exit code 1 on stderr.

List and page

list defaults to active. Use --status resolved or --status all for the other views. Output is { data, total }: one page of incidents and the number matching the filter. --limit accepts 1 to 100. Increase --offset to fetch later pages.

Acknowledge and resolve

API keys do not identify a member, so acknowledgment requires --by with a member UUID in the account. Acknowledgment stops escalation. A resolved or already acknowledged incident cannot be acknowledged.

Resolving a published incident always posts a public resolution and emails its status-page subscribers by default. Add --no-notify-subscribers to suppress that email. --message supplies the public resolution text.

Change service or metadata

update changes only the title, severity and service fields you supply. It does not set lifecycle status. Severity is minor, major or critical.

npx siteqwality incidents update "$INCIDENT_ID" --service "$SERVICE_ID"
npx siteqwality incidents update "$INCIDENT_ID" --clear-service

Moving an incident to a service can start its escalation policy. --clear-service detaches it. Choose either flag, not both.

Failed requests

All commands accept --api-url, --timeout and --retries. Reads retry transient failures within the configured budget. Incident mutations make one request even with --retries, so an ambiguous timeout cannot automatically repeat paging or a public update. Inspect the incident before retrying a failed mutation.

The CLI does not follow redirects, accepts HTTPS (or loopback HTTP for local tests), and keeps API keys and response bodies out of error messages.