Nerd Guild

Service Status

Operational health is tracked across authentication, API availability, import processing, and data storage. This page summarizes the forward-looking monitoring baseline for launch.

Request volume

24h traffic

Track by route, environment, and auth status

Latency

p95 and p99

Review API, Lambda, DynamoDB, and S3-based flows

Errors

5xx and throttles

Alarm on sustained spikes per service

Authentication

sign-in success rate

Alert on Cognito login and token refresh failures

Import processing

PDF and upload queue health

Watch failures, retries, and queue build-up

Storage growth

DynamoDB and S3 usage

Forecast churn and unexpected data growth

Current service posture

  • Authentication health is reviewed against sign-in success rate and token refresh failures.
  • API health is reviewed across 5xx responses, latency, and Lambda throttling.
  • Import and file workflows are monitored for queue buildup, retries, and payload failures.
  • Storage usage is reviewed for budget drift, retention health, and unexpected growth.

Operations runbooks

  1. Login outage

    Verify Cognito status, confirm app/client configuration, check session cookie expiry, and restore sign-in using the latest known-good config.

  2. API outage

    Check CloudWatch, API Gateway stage health, Lambda errors, throttle alarms, and dependency availability before rolling back or draining traffic.

  3. Corrupt import

    Quarantine the failing import, review parser logs, validate file type/size constraints, and rerun only the verified artifact after a safe restore.

  4. Runaway cost

    Review budgets, adjust concurrency or retention, identify the resource driving spend, and limit or disable the offending workflow until the blast radius is contained.

  5. Data restore

    Freeze writes, restore the last known-good backup or point-in-time snapshot, reconcile integrity checks, and validate reads and writes before resuming traffic.

  6. Compromised credential

    Rotate the affected secret or app client, revoke active sessions, review CloudTrail and Cognito activity, and validate the replacement path in staging before production re-entry.

Incident severity

  • SEV-1

    Critical user-impacting service outage or compromise

    Acknowledge within 15 minutes and notify owner immediately

  • SEV-2

    Major but contained degradation affecting a core workflow

    Acknowledge within 1 hour and maintain an incident bridge

  • SEV-3

    Degraded performance or limited user impact with clear mitigation

    Acknowledge within 4 hours and track corrective action

  • SEV-4

    Low-impact issue or support request with no data exposure

    Acknowledge within 1 business day

Incident notification contacts

  • Incident commander / final escalation

    Kit Cloud / project owner

    support@nerdguild.org

  • Billing and cost overrun escalation

    Kit Cloud / project owner

    support@nerdguild.org

  • Security and abuse reports

    Kit Cloud / project owner

    support@nerdguild.org (see SECURITY.md)

Recurring operational reviews

  • Dependency review (Every release)

    Run `npm audit --audit-level=high` and review outdated high-risk packages before promoting a release.

  • Access review (Quarterly)

    Review IAM roles, Cognito app clients, and any individuals with AWS console or CLI access.

  • Cost review (Monthly)

    Compare actual AWS spend against the budget target and investigate any sustained variance.

  • Backup review (Quarterly)

    Confirm DynamoDB point-in-time recovery and S3 versioning remain enabled across environments.

  • Recovery drill (Before each production launch gate and at least annually after)

    Execute the staging restore procedure in RESTORE.md and record the outcome.