Vendor Incident Coordination

When the vendor is the incident, your team coordinates with them. The pattern.

Escalate to vendor

Vendor escalation is its own discipline. Know the path before the incident, not at 3 AM when the support form is asking you to log in. Premium support tiers pay back in saved minutes during serious incidents; documented contacts and SLA references mean nobody asks "who do we even call?"

• Account manager or premium support. Named escalation path per vendor. Faster than the generic support queue.
• Premium tier pays back. Saved minutes during serious incidents justify the recurring premium cost.
• Documented contact per vendor. Named contact and after-hours number captured per vendor. Catches "who do we call?" panic.
• SLA reference per vendor. Contractual response time documented. Supports holding vendors to their commitments.

Comms

Comms during a vendor incident borrow the vendor's narrative. Your status page references the vendor's status page rather than duplicating updates; customer-impact framing translates vendor language into terms your customers actually understand. Post-incident, you hold the vendor to their postmortem commitments.

• Vendor status page is source of truth. Canonical incident page per vendor. Link rather than paraphrase.
• Your status page references vendor. Upstream link per incident. Avoids duplicating vendor-side updates that may diverge.
• Customer-impact framing. "What does this mean for you" translation per incident. Vendor language often does not fit your customers.
• Post-incident vendor postmortem. Expectation set per vendor. Holds vendors accountable to their public commitments.

Recover

Recovery is not automatic when the vendor recovers. Verify your systems also recovered; some integrations need explicit re-sync (webhooks, queues, cached state); per-integration runbook captures the re-sync command so the on-call does not have to derive it. Extended monitor window after recovery catches latent issues.

• Verify your systems recovered. Post-vendor-recovery check per incident. Vendor recovery does not always mean your stack is fine.
• Some integrations need re-sync. Explicit reconciliation per integration. Webhooks, queues, cached state may need attention.
• Documented recovery step per integration. Runbook re-sync command per integration. Catches half-recovered state.
• Post-recovery monitor window. Extended watch per incident. Catches latent vendor-side issues that surface after the all-clear.