Production Go-Live Checklist
A 130-point go-live checklist across governance, architecture, security, data, performance, observability, deployment, disaster recovery, compliance and cutover day.
17-23 minutes(3453 words)complex
Quick Navigation
Difficulty: Advanced
Estimated Time: 35-50 minutes
Prerequisites: Production operations experience, Familiarity with ITIL change management, Understanding of backup and recovery, Basic security concepts
A go-live checklist is only useful if it can block a release. This one assigns every item a criticality and an owner so the go/no-go decision is mechanical rather than a matter of who is most confident in the room.
Criticality: B = Blocking, no GO without it. M = Major, GO possible with a dated action plan. m = Minor, can be handled post-production.
Owners: ARCH (architect), DEV, OPS/SRE, DBA, SEC (security), NET (network), QA, PO (product owner), SUP (support tier 1/2), LEG (legal/DPO).
1. Governance and Project Prerequisites
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| G01 | Go-live scope frozen | Signed list of delivered features plus exclusions | PO | B |
| G02 | GO/NO-GO decision scheduled | Cutover committee dated, participants and quorum defined | PO | B |
| G03 | GO/NO-GO criteria written down | Objective thresholds (blocking bugs = 0, performance under X ms) | PO/QA | B |
| G04 | Cutover window approved | Slot approved by the business and outside critical periods | PO | B |
| G05 | User communication sent | Email or intranet at D-7 and D-1 announcing downtime | PO | M |
| G06 | Change freeze in effect | Freeze declared on all impacted components | OPS | M |
| G07 | Change request approved (ITIL) | CAB ticket approved, window and impacts documented | OPS | B |
| G08 | Run budget and licences approved | Monthly infrastructure and production licence cost signed off | PO | M |
| G09 | Supplier contracts and SLAs active | Vendor, hosting and CDN support active in production, not trial | PO | B |
| G10 | Crisis team identified | Named list with phone numbers and a dedicated war room channel | OPS | B |
2. Architecture and Design
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| A01 | Technical architecture document current | Flow diagrams, components, ports and protocols validated | ARCH | B |
| A02 | No SPOF identified | Documented single-point-of-failure analysis | ARCH | B |
| A03 | Multi-AZ or multi-site redundancy | Instances across two or more zones, failover tested | ARCH/OPS | B |
| A04 | Sizing validated | CPU, RAM, IOPS and storage calculated on expected peak plus 30 percent margin | ARCH | B |
| A05 | Scalability strategy defined | Autoscaling configured with min/max/thresholds, or a manual procedure | OPS | M |
| A06 | External dependencies mapped | Third-party API list with SLAs and behaviour when unavailable | ARCH | B |
| A07 | Degraded modes defined | Circuit breaker, fallback and emergency cache documented | ARCH/DEV | M |
| A08 | Version compatibility matrix | OS, runtime, DBMS and middleware on supported, non-EOL versions | ARCH | B |
| A09 | Batch and API idempotency | Replay without duplication tested | DEV | M |
| A10 | Timeouts and retries handled | Values defined per flow, with exponential backoff | DEV | M |
3. Environments and Configuration
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| E01 | Staging matches production | Deviations documented and accepted | OPS | M |
| E02 | Infrastructure as Code | Terraform or Ansible versioned, plan shows no drift | OPS | M |
| E03 | Strict environment separation | No cross access from dev to production (network, accounts, data) | SEC | B |
| E04 | Configuration externalised | No environment value hardcoded in the source | DEV | B |
| E05 | Secrets in a vault | Vault, KMS or Secret Manager, no secret in Git or a flat file | SEC | B |
| E06 | Secret rotation planned | Procedure and frequency defined | SEC | M |
| E07 | NTP synchronisation | All servers on a common time source, drift under 1 s | OPS | M |
| E08 | Consistent timezones and locales | UTC in the database, converted at presentation | DEV | M |
| E09 | System quotas and limits | ulimit, file descriptors, max connections and tmp size verified | OPS | M |
| E10 | Disk space and log rotation | Alert threshold at 80 percent, logrotate active, purge tested | OPS | B |
4. Security
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| S01 | Penetration test completed | Report delivered, critical and high vulnerabilities fixed | SEC | B |
| S02 | Infrastructure and image vulnerability scan | Zero critical CVEs, high CVEs tracked with a deadline | SEC | B |
| S03 | SAST/DAST/SCA analysis of the code | CI report green, no dependency with a critical flaw | DEV/SEC | B |
| S04 | Server hardening | CIS baseline applied, unnecessary services disabled | OPS/SEC | M |
| S05 | Encryption in transit | TLS 1.2 or above enforced, HSTS on, weak suites disabled | SEC | B |
| S06 | Encryption at rest | Disks, databases and backups encrypted with managed keys | SEC | B |
| S07 | Certificates valid and monitored | Validity over 90 days, expiry alerting, automatic renewal | OPS | B |
| S08 | Identity and access management | Named accounts, MFA, least privilege | SEC | B |
| S09 | Default accounts removed or renamed | No default passwords, test accounts purged | SEC | B |
| S10 | Dedicated service accounts | One technical account per service, non-human, non-interactive | SEC | M |
| S11 | WAF, anti-DDoS and rate limiting | Rules active and tested, thresholds per IP and per user | SEC/NET | M |
| S12 | Security logging to SIEM | Auth events, failures, privilege changes and sensitive data access collected | SEC | M |
| S13 | Data anonymised outside production | No real production data in staging or dev | SEC/DBA | B |
| S14 | HTTP security headers | CSP, X-Frame-Options, X-Content-Type-Options, Referrer-Policy | DEV | M |
| S15 | No sensitive data in logs | Passwords, tokens, card numbers and health data masked | DEV | B |
| S16 | Access revocation procedure | Departure or incident: revocation within 24 h, documented | SEC | M |
5. Network, DNS and Exposure
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| N01 | Flow matrix validated and applied | Source, destination, port, protocol; firewall rules in place | NET | B |
| N02 | Unnecessary ports closed | External and internal scan confirming minimal exposure | SEC/NET | B |
| N03 | DNS records created | A, CNAME, MX and TXT correct, TTL lowered before cutover | NET | B |
| N04 | Load balancer and health checks | Application-level probes, not just TCP, thresholds validated | OPS | B |
| N05 | Load distribution tested | Removing a node causes no service interruption, test performed | OPS | M |
| N06 | CDN and cache configured | Cache rules set, purge tested, headers consistent | OPS | m |
| N07 | VPN and remote access secured | Bastion or VPN mandatory for administration | SEC/NET | B |
| N08 | Inter-site bandwidth and latency | Measurements meet application requirements | NET | M |
| N09 | Fixed IPs and partner allowlists | Egress IPs communicated to and authorised by third parties | NET | M |
6. Data and Database
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| D01 | Migration scripts tested | Run against a production copy, duration measured | DBA/DEV | B |
| D02 | Database rollback script | Tested and timed, or a validated restore strategy | DBA | B |
| D03 | Full backup before cutover | Backup verified and timestamped, restore point known | DBA | B |
| D04 | Restore tested, not just the backup | Real restore test under RTO, integrity checked | DBA | B |
| D05 | Retention policy defined | Retention periods per data type, purge automated | DBA/LEG | M |
| D06 | Indexes and query plans optimised | Slow queries identified and fixed | DBA | M |
| D07 | Database replication and HA | Replica active, lag monitored, failover tested | DBA | B |
| D08 | Initial data migration | Volume, consistency and reconciliation validated by the business | PO/DBA | B |
| D09 | Character sets and encoding | UTF-8 end to end, tested with accents and special characters | DEV/DBA | M |
| D10 | Statistics and maintenance scheduled | VACUUM, ANALYZE, reindex or equivalent scheduled | DBA | m |
7. Application and Quality
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| Q01 | User acceptance testing signed off | Acceptance record signed by the business | PO/QA | B |
| Q02 | No open blocking defects | Defect table: zero blocking, majors triaged | QA | B |
| Q03 | Regression tests passing | Automated suite green, coverage rate known | QA | B |
| Q04 | Inter-application integration tests | All partner flows tested under real conditions | QA | B |
| Q05 | End-to-end tests of critical journeys | Top five business journeys validated | QA/PO | B |
| Q06 | Browser and device compatibility | Target matrix tested across desktop, mobile and versions | QA | M |
| Q07 | Accessibility (WCAG) | Target level reached, or gaps documented | DEV | m |
| Q08 | User error handling | Custom 404 and 500 pages, no stack trace exposed | DEV | M |
| Q09 | Versioning and release tag | Git tag frozen, binary or image immutable and signed | DEV | B |
| Q10 | Feature flags under control | State of every flag in production documented and reversible | DEV | m |
8. Performance and Capacity
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| P01 | Nominal load test | Target: X concurrent users, response time under threshold | QA/OPS | B |
| P02 | Stress test | Behaviour beyond target load known, breaking point identified | QA | M |
| P03 | Soak test | 8 to 24 h with no memory leak or degradation | QA | M |
| P04 | P95 and P99 response times measured | Baseline values recorded for post-production comparison | OPS | M |
| P05 | Resource consumption under load | CPU under 70 percent, RAM under 80 percent, IOPS under threshold | OPS | M |
| P06 | Connection pools sized | Database, HTTP and queue pool sizes consistent with load | DEV/DBA | M |
| P07 | Application cache validated | Hit rate measured, invalidation strategy tested | DEV | m |
| P08 | 12-month storage capacity projected | Estimated growth plus alert thresholds | OPS | M |
9. Observability
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| O01 | Infrastructure monitoring active | CPU, RAM, disk, network and processes on every node | OPS | B |
| O02 | Application monitoring (APM) | Traces, errors and latency per endpoint | OPS/DEV | B |
| O03 | Business monitoring | Business probes such as orders per minute or payment failures | PO/OPS | M |
| O04 | Centralised logging | Every component ships to the log platform | OPS | B |
| O05 | Correlated logs (trace ID) | Correlation identifier propagated end to end | DEV | M |
| O06 | Alerts defined with thresholds | Every alert has a threshold, a recipient and an action | OPS | B |
| O07 | Alerts tested by real firing | End-to-end test: trigger through to notification received | OPS | B |
| O08 | On-call and escalation configured | Rotation, reachability, escalation from tier 1 to 3 | OPS/SUP | B |
| O09 | Operational dashboards | A service health dashboard accessible to the team | OPS | M |
| O10 | Health and readiness endpoints | /health and /ready distinct and not publicly exposed | DEV | M |
| O11 | Job and batch monitoring | Alert on failed, late or skipped jobs | OPS | M |
| O12 | Alert noise controlled | No non-actionable alerts, grouping rules in place | OPS | m |
10. Deployment and CI/CD
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| C01 | Automated deployment pipeline | Reproducible deployment with no manual intervention | OPS | M |
| C02 | Deployment tested in pre-production | Same pipeline and same artifact as production | OPS | B |
| C03 | Deployment strategy defined | Blue/green, canary or rolling, choice documented | ARCH/OPS | M |
| C04 | Rollback procedure written and tested | Rollback under X minutes, tested at least once | OPS | B |
| C05 | Component start order | Startup and shutdown sequence documented with dependencies | OPS | M |
| C06 | Graceful shutdown | Connections drained, no transaction lost | DEV | M |
| C07 | Artifacts archived and traceable | Deployed version identifiable from the app via /version | DEV | m |
| C08 | Deployment rights restricted | Only authorised people can deploy to production | SEC | B |
11. Resilience, Backup and Disaster Recovery
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| R01 | RTO and RPO defined and approved | Values agreed with the business | PO/ARCH | B |
| R02 | Backups automated and scheduled | Frequency meets the RPO, success monitored | OPS | B |
| R03 | Off-site backups (3-2-1 rule) | Copy outside the site or cloud account, immutable if possible | OPS | B |
| R04 | Restore test documented | Dated test record less than 6 months old | OPS | B |
| R05 | Disaster recovery plan written | Scenarios, sequence, owners and estimated duration | OPS | M |
| R06 | Failover test performed | Real or simulated failover with a report | OPS | M |
| R07 | Dependency failure resistance | Third-party service outage produces the correct degraded mode | QA/DEV | M |
| R08 | Queues and replay | Dead letter queue configured, replay procedure written | DEV/OPS | M |
12. Compliance, Legal and Personal Data
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| J01 | Processing register updated | Processing declared, legal basis identified | LEG | B |
| J02 | DPIA completed where required | Impact assessment validated by the DPO | LEG | B |
| J03 | Data localisation compliant | Hosting within the authorised geographic zone | LEG/OPS | B |
| J04 | Data subject rights implemented | Access, rectification, erasure and portability operational | DEV/LEG | M |
| J05 | Privacy policy and terms published | Texts online and legally approved | LEG | M |
| J06 | Cookie banner and consent | Compliant, with refusal as easy as acceptance | DEV/LEG | M |
| J07 | Sector compliance | PCI-DSS, HDS, ISO 27001, DORA as applicable | SEC/LEG | B |
| J08 | Processor agreements signed | GDPR clauses with every host and vendor | LEG | M |
| J09 | Traceability and legal archiving | Audit logs retained for the regulatory period | SEC | M |
13. Documentation and Handover
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| K01 | Operations runbook | Start, stop, restart, purge and routine procedures | OPS | B |
| K02 | Troubleshooting guide | Top ten likely incidents with resolutions | OPS/DEV | M |
| K03 | Installation and deployment documentation | Reproducible by a third party | OPS | M |
| K04 | Flow and dependency map | Diagram current and accessible | ARCH | M |
| K05 | API documentation (OpenAPI) | Published and current for consumers | DEV | M |
| K06 | Tier 1 and 2 support trained | Session delivered, training material handed over | SUP | B |
| K07 | End-user training | Sessions, webinars or guides distributed | PO | M |
| K08 | Knowledge base populated | KB articles created for frequent cases | SUP | m |
| K09 | Operations RACI matrix | Who does what in run, clearly established | OPS/PO | M |
| K10 | Contact details for all parties | Crisis directory current, internal and vendors | OPS | B |
14. Cutover Day
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| B01 | Minute-by-minute cutover plan | Each task with time, duration, owner and prerequisites | OPS | B |
| B02 | Points of no return identified | Rollback decision deadline defined | OPS/PO | B |
| B03 | Pre-cutover backup and snapshot | Taken and verified immediately before the operation | OPS/DBA | B |
| B04 | Maintenance page ready | Tested, message approved, fast to enable and disable | OPS | M |
| B05 | Inbound flows and batches stopped | Scheduler suspended, partners notified | OPS | M |
| B06 | War room open | Channel, video call and participants connected at H-1 | OPS | M |
| B07 | Post-cutover smoke tests | Short checklist: login, key journey, inbound and outbound flow | QA | B |
| B08 | Business validation before reopening | PO confirms correct operation before public opening | PO | B |
| B09 | Batches and flows reactivated | Scheduler restarted, first run monitored | OPS | B |
| B10 | End-of-operation communication | Service restored message to users and support | PO/SUP | M |
15. Post-Production (D+1 to D+30)
| ID | Control point | Acceptance criteria / evidence | Owner | Crit. |
|---|---|---|---|---|
| H01 | Hypercare period | Duration defined, for example two weeks, team mobilised | OPS/SUP | M |
| H02 | Daily indicator review | Errors, latency and volume compared to baseline | OPS | M |
| H03 | User ticket tracking | Daily triage, trend monitored, recurring themes addressed | SUP | M |
| H04 | First production backup verified | Real backup executed and restore tested | OPS | B |
| H05 | First monthly batch or close verified | Execution monitored, results validated by the business | PO/OPS | M |
| H06 | Legacy system decommissioned | Planned after a safety period, data archived | OPS/PO | M |
| H07 | Retrospective | Meeting held, improvement actions tracked | PO/OPS | m |
| H08 | Formal handover to the run team | Project-to-operations transfer record signed | OPS | M |
The Decision Rule
GO if 100 percent of B items are validated and every open M item has a dated action plan with a named owner.
The rule matters more than the list. A checklist without a decision rule becomes a document people acknowledge rather than a gate that stops a bad release, and the items most likely to be quietly skipped are D04 (restore actually tested) and C04 (rollback actually tested) — the two that only pay out on the worst day.
Tags: #GoLive #ProductionReadiness #SRE #DisasterRecovery #Checklist