Ecosystem Climate Capital /Technical Documentation

Operations

Standard Procedures

The operating procedures every team member follows. They make outcomes reproducible, auditable, and safe.

SOP-001 — Development workflow

  1. Pick a ticket from the current sprint; move it to In progress.
  2. Create a branch feat/<ticket-id>-short-slug.
  3. Write tests first when modifying domain logic; cover edges.
  4. Open a draft PR early; CI must pass (lint, typecheck, unit, e2e on critical paths).
  5. Request review from a domain owner; at least one approval required.
  6. Squash-merge with a Conventional Commit message; preview deploys auto-publish.

SOP-002 — Release

  1. Releases ship every weekday at 14:00 UTC; hotfixes any time.
  2. Release captain rotates weekly; they own the release notes and rollback decision.
  3. Tag vYYYY.MM.DD.N; staging soaks for 30 min before promotion.
  4. Production deploy is gated by a green synthetic monitor run.
  5. Post-release: monitor RED metrics dashboard for 30 min; close release in changelog.

SOP-003 — Incident response

  1. Anyone can declare an incident in #incidents with severity.
  2. Incident commander assigned within 5 min for SEV-1 / SEV-2.
  3. Communications lead posts customer updates every 30 min on the status page.
  4. Mitigate first, fix later; rollback is always an acceptable first step.
  5. Postmortem within 5 business days; blameless, action items tracked to completion.
SeverityDefinitionResponse
SEV-1Platform down or data lossPage on-call, all-hands
SEV-2Major feature broken for > 5 % usersPage on-call
SEV-3Degraded, workaround existsNext business day

SOP-004 — Data change management

  1. Production data changes require a ticket and a peer-reviewed migration script.
  2. Run on staging first; capture before/after row counts.
  3. Apply with a transaction; have rollback SQL ready.
  4. Log the change in the data-ops journal with diff and approver.

SOP-005 — MRV sign-off

  1. Measurements are imported and auto-enriched with satellite layers.
  2. Independent verifier reviews evidence; rejects with reason or approves.
  3. Approved batches are hash-stamped; the hash is anchored to the audit log.
  4. Registry export reviewed by methodology owner before pushing.
  5. Issuance recorded in marketplace inventory only after registry confirms.

SOP-006 — Access management

  1. Access is granted by role, not by individual permissions.
  2. New hires onboarded via the access-request workflow; manager approves.
  3. Quarterly access review by tenant admin; revocations applied within 24 h.
  4. Offboarding triggers immediate session invalidation and device wipe.

SOP-007 — Backups & disaster recovery

  1. Continuous WAL shipping to a separate region.
  2. Nightly snapshots retained for 35 days; monthly for 12 months.
  3. Quarterly restore drill against an isolated environment; result published internally.