- Which guide should I start with?
- Get started, then work through Set it up in order. Each step depends on the one before it — channels before escalation, escalation before on-call, alerting last.
- Why is alerting the last setup step?
- An alert that fires before there is a workspace, a channel, an escalation level and someone on call has nowhere to go. Connecting the source last means the first real alert lands on a path that has already been tested.
- What is the difference between a workspace and a team?
- A workspace groups the people and services in the same incident-response flow. A team is the group inside a workspace that escalation actually addresses — levels point at teams.
- What is the difference between Admin and Member?
- Admins have read and write access to Shankh configuration. Members have read access plus the ability to acknowledge and resolve incidents, so they can still participate fully in incident response.
- Which notification channels can I use?
- Slack, Telegram and mobile push. WhatsApp, voice calls, SMS and email are not currently available for incident delivery, even where they appear in the interface.
- Which clouds does Shankh support natively?
- AWS has native discovery and alert setup. Grafana, Jenkins and SigNoz connect as Vörr connectors. Anything else can send to the organisation's private webhook.
- Why does acknowledging matter so much?
- Acknowledgement is how Shankh knows the incident has been picked up. Without it escalation continues to other people, and response-time reporting has nothing to measure.
- What happens if nobody is on call?
- If no Level 0 on-call person is available when an incident is processed, Shankh can move to the next configured escalation level. Treat that as a backstop, not as coverage.
- Can Shankh analyse an incident with AI?
- Not today. AI incident analysis is not implemented — the AI summary and AI postmortem in Reports are text templates assembled from the incident record, with no model call. Alerting, routing, escalation, acknowledgement, resolution and reporting are unaffected.
- Who can see a stakeholder live page?
- Anyone holding the link. Keep internal hostnames, customer names and credentials out of incident titles, descriptions and progress updates.