Incident management

Incidents are moments of high stress for our clients. Our job is to take control quickly, restore service as fast as we can, and keep everyone informed while we work. When something looks like an incident, treat it like one until you are confident it is resolved.

Ownership

Incidents must have one clearly named owner at all times. The owner:

• Leads the response, even if others are doing the hands-on work.
• Handles all outward communication and status updates.
• Makes sure we never “drop the ball”—if you own it, you stay on it until the baton is deliberately passed to someone else.

If you cannot continue owning the incident (e.g. your day ends, you have to join another call, you are about to commute), arrange a handover before stepping away. Share the current status, outstanding actions, and the agreed timeline for future updates, and confirm the new owner acknowledges the handover.

Communication

Updates should be visible to the client in the agreed channel (usually email or Slack). Send an update as soon as you take ownership and maintain a predictable cadence afterwards. Each update must:

• Acknowledge the current impact, even if nothing has changed yet.
• Describe the steps we have taken since the previous update.
• Outline what we are doing next.
• Commit to the exact time of the next update—and meet it.

If we are still investigating, say so plainly and explain what we are looking into. Progress notes are better than silence. When we are blocked waiting for a third party, tell the client who we are waiting on and what we expect to happen next.

Regular updates

Before finishing each update, decide when the next one will be. Pick a cadence that reflects the severity of the incident (e.g. every hour). Put that commitment in writing so the client knows when to expect to hear from us again. If new information means you need to change the scheduled time, communicate the change immediately and ensure someone is actively monitoring the clock.

Presence

If you need to pause, let the client know in advance and make sure another engineer is actively covering the incident. The handover should include:

• The latest update sent to the client.
• What we are doing next and who owns each action.
• The time of the next client update.

Only step away once the incoming owner has confirmed they are taking over and will send the next update. If no one can cover, tell the client explicitly when we will be back online and what they should expect during that gap.

Closing

When the issue is resolved, the owner sends the wrap-up message summarising:

• What happened and when service was restored.
• Any customer impact or follow-up actions.
• What we are doing to prevent a recurrence (or when to expect that analysis).

Capture a brief internal log—including timestamps of ownership changes and client updates—to make the incident review smoother.