Prepare and test your skills
Prepare and test your skills
Analyzing and defining technical processes is the work of making software delivery fast, safe, and repeatable on Google Cloud. Code moves from commit to production as one immutable artifact with security checks, validation, observability, governed provisioning, and recovery built in. Each process trades cost, risk, and speed against the others.
Software development lifecycle (SDLC) on Google Cloud is the automated path that moves code from source commit to production with checks at every stage. Cloud Build compiles and tests code, Artifact Registry (opens in a new tab) stores versioned container images, and Cloud Deploy promotes the same image through development, staging, and production. Traffic flows from the source repository to Cloud Build, then from Cloud Build to Artifact Registry, then from Cloud Deploy to each target environment.
A left-to-right process flow from source commit through Cloud Build, Artifact Registry, Binary Authorization gate, and Cloud Deploy promoting the same image digest to development, staging, and production.
The enterprise sequence runs in order: first the commit triggers the build, then Cloud Build creates the image and pushes it to Artifact Registry, then Cloud Build runs unit tests and Container Structure Tests, then Artifact Analysis scans the image for vulnerabilities, and only after approval Binary Authorization signs the image with an attestation. The same image digest moves across environments and is never rebuilt, while configuration and secrets are injected at deploy time from Secret Manager. Release images use semantic versioning tags, while pre-release images use the Git commit hash that ties the image to the exact change.
Deployment strategy evaluation is the choice of how a new version replaces the old version to limit disruption and keep a clear rollback path. Google Kubernetes Engine (GKE) natively supports recreate and rolling strategies, while blue-green on GKE requires additional tooling such as Spinnaker. Cloud Deploy adds gated approvals so promotion to a target only proceeds after required approval.
| Strategy | Traffic shift | Resource and compatibility requirement |
|---|---|---|
| Blue-green | Traffic flows from the old server set to the new server set in one switch after verification | Temporarily requires up to twice the production servers; database schemas must stay backward compatible |
| Rolling | Traffic flows gradually from old servers to new servers as each server is updated | Updates servers in place without a duplicate fleet; requires load balancer behavior that avoids bouncing users between versions |
Choose blue-green when avoiding interruption during cutover outweighs the cost of duplicate capacity. Choose rolling when the goal is early detection with a smaller capacity footprint. Shift-left security supports both choices by running builds in isolated and ephemeral environments with granular Identity and Access Management permissions and Supply-chain Levels for Software Artifacts (SLSA) Level 3 provenance.
Exam tip: Artifact Analysis finds and tracks vulnerabilities in images, while Binary Authorization blocks deployment unless the required attestations prove the checks passed.
Continuous integration and continuous deployment is the automation that builds every change, stores one immutable artifact, and promotes that artifact safely. Cloud Deploy defines the ordered progression of targets, a release represents a version to deploy, and a rollout executes that release into a specific target. An Automation is a child resource of one delivery pipeline that stores the target selector and rules, while an AutomationRun is created each time the automation triggers.
Cloud Deploy supports four automated actions:
The workflow runs in order: first configure the Automation, then register it, then create a release that starts rollouts in order, then Cloud Deploy triggers promote or advance logic on success or repair logic on failure. You can automate up to 250 automation rules per delivery pipeline.
Troubleshooting and root cause analysis is the systematic practice of separating network failures from permission failures and application failures from platform incidents. The method first checks whether traffic can travel from source to destination, then checks whether the identity is allowed to perform the action. Network checks use configuration analysis and flow logs, while permission checks use Policy Troubleshooter.
Connectivity Tests (opens in a new tab) in Network Intelligence Center analyzes the configured path from a source endpoint to a destination endpoint without sending live traffic. Traffic flows from the source, such as a virtual machine (VM) or subnet, across Virtual Private Cloud (VPC) routes, peering connections, and virtual private network (VPN) tunnels to the destination. VPC Flow Logs records actual sampled traffic entering and leaving VMs per subnet and sends entries to Cloud Logging, while Firewall Rules Logging shows which rule allowed or denied traffic. Cloud Audit Logs records who did what, where, and when, with Admin Activity logs always enabled and Data Access logs disabled by default except for BigQuery.
Systematic diagnosis uses reliability goals and observability tools to detect, isolate, and resolve failures. Reliability starts with service level objectives (SLOs) such as 99.9% API success and 95th percentile latency below 300 ms. Cloud Monitoring analyzes performance data that flows from infrastructure and applications into dashboards and alerts, including the four golden signals: latency, traffic, error rate, and saturation of CPU, memory, GPU, or TPU.
Cloud Logging links a failed request to its timestamp, service, stack trace, and recent change, while Cloud Trace follows a single request from frontend through each downstream microservice to pinpoint which span adds latency. Error Reporting groups repeated crashes by signature and Cloud Profiler reveals which functions consume time or memory. The runbook moves in order: first communication notifies stakeholders, then triage determines scope, then investigation separates symptoms from cause, then resolution applies the fix. A blameless post-mortem explains what happened and what will change without blaming individuals, and progressive deployment expands verified changes gradually.
Testing and validation is the process of checking code, templates, performance, and recovery before and after deployment. Validation happens before provisioning to block misconfigurations, during CI/CD to block untested artifacts, and after deployment to detect drift. Together these layers prevent expensive production fixes.
| Method | What it checks | When it runs |
|---|---|---|
| Linting | Syntax, formatting, and basic configuration errors | First, as soon as the template is written or committed |
| Policy-as-code enforcement | Allowed regions, encryption, identity controls, and other security rules | Second, by evaluating the template against codified policies |
| Pre-deployment simulation | Which resources would be created, modified, or deleted | Third, by previewing the plan before provisioning |
Performance testing uses load testing for normal and peak traffic and stress testing to find bottlenecks, with Cloud Load Balancing distributing simulated traffic to test targets. Capacity planning translates forecast users, request rate, and growth into managed instance group size, machine type, and disk count with N+1 redundancy. Disaster recovery validation proves that backup frequency meets recovery time objective (RTO) and recovery point objective (RPO), where RTO is how long until service resumes and RPO is how much data in writes can be lost.
Service catalog and provisioning is the governed self-service model that lets teams deploy quickly from approved choices without bypassing controls. Service Catalog provides this control under a Google Cloud organization, where the organization holds the catalog, the catalog holds the solutions, and sharing grants visibility to users in specific folders and projects. A Deployment Manager template enforces region and machine type limits at launch, while a reference link distributes only admin-reviewed documentation.
The lifecycle moves through ordered states: first request with owner and cost center, then placement and approval to folder and billing account with quota checks, then provisioning with standard APIs and labels, then active use within Shared VPC and budget alerts, then decommissioning when the project is no longer needed. HashiCorp Terraform defines desired state in files, terraform plan previews changes, and terraform apply creates infrastructure and writes a state file. The google_*_iam_member resource adds members without claiming the full policy, while google_*_iam_policy and google_*_iam_binding are authoritative and overwrite all permissions.
Disaster recovery on Google Cloud turns business limits for downtime and data loss into replication, backup, and traffic failover across regions. Smaller RTO and RPO require more replication and standby capacity, so cost rises as objectives approach zero. Turbo replication replicates to both regions of a dual-region bucket with a 15-minute RPO, while Firestore regional databases replicate across three zones for 99.99% availability and multi-region databases replicate across five zones in three regions for 99.999% availability.
| Pattern | Recovery readiness | Cost and use |
|---|---|---|
| Cold standby | No running servers; images and backups wait for failover | Lowest cost; largest RTO and RPO; fits batch jobs |
| Warm standby | Minimal suite runs, such as small database replica with async replication | Medium cost; smaller RTO and RPO |
| Hot standby | Both environments serve production with replication | Highest cost; lowest RTO and RPO; fits Tier 1 payments |
Traffic management shifts clients automatically when health checks fail. The primary is the global external Application Load Balancer, while backups are regional external Application Load Balancers. Cloud DNS controls failover from primary to backups and failback after recovery. Business-critical recovery must depend only on data plane components that continuously process operations, not on management plane operations such as VM-creation APIs.
Continue with this study guide: Analyzing and defining technical processes
Analyzing and defining business processes is the work of connecting cloud architecture to cost, risk, skills, and measurable business value. Stakeholders agree on tradeoffs from shared facts, changes ship gradually with guardrails, teams build required skills, and spend and health data guide decisions. Governance keeps self-service fast without losing control.
Stakeholder management is the practice of aligning technical and non-technical owners around cloud decisions through shared facts and facilitated agreement. Product owners, developers, administrators, and financial analysts weigh cost, speed, risk, and performance differently, so the architect makes tradeoffs explicit. Consensus holds when everyone works from the same trusted data about cost mapped to key performance indicators (KPIs).
Defining platform roles clarifies who owns strategy, delivery, security, and assurance in a distributed platform such as a data mesh (opens in a new tab). The Data platform product owner owns strategy and roadmap, the Data platform engineer owns reusable delivery, the Platform and security engineer owns enterprise fit for networking and security, and the Enterprise architect owns alignment through design authority. Influence flows from leadership into the product owner, from the architect outward for assurance, and from central engineers to producers and consumers through shared templates. Workshops turn business needs into self-service templates for Terraform, pipelines, and containers, bounded by shared responsibility and location rules such as data residency.
Change management is the system of culture, pipelines, and controls that lets teams release often without long outages. Site Reliability Engineering (SRE) balances feature velocity with availability and latency, while DevOps connects development and operations through shared tooling. Both sides use the same deployment pipeline, automation, runbooks, and reviews, and applications are designed for testability with disposable test environments.
Source code flows from Cloud Storage, GitHub, or Bitbucket into Cloud Build for build and test, into Artifact Registry for versioned storage, and into Cloud Deploy for ordered promotion across development, test, and production. Automated tests act as gates, so higher coverage allows more changes before reliability drops. Risk is controlled with Identity and Access Management least privilege, Organization Policy guardrails, and progressive rollout in order: first pilot in one zone, then revert on issue, then diagnose outside production, then expand gradually. Cloud Audit Logs records who did what, where, and when for every governance change.
Team assessment and skills readiness is the process of defining needed cloud roles, measuring current ability, and closing gaps before go-live. Key roles typically include Cloud Architects, Data Engineers, and Security Specialists, plus operators for deployment and incident response. The assessment checks whether staff can design workloads, store images in Artifact Registry, scan and sign code, deploy to targets, and inject secrets and keys at runtime.
Ready teams secure the artifact supply chain with Artifact Analysis for scanning, Software Bill of Materials (SBOM) generation, and Assured Open Source Software (Assured OSS) for validated packages. Controls include cleanup policies, virtual repositories that prefer internal sources, immutable tags or deployment by SHA-256 digest, and Cloud Build quality gates. Binary Authorization verifies container integrity at deploy time with progressive gates and monitored breakglass bypass. Culture sustains skills through psychological safety, blameless learning, and Start-Stop-Continue retrospectives that name what to start, stop, and continue.
Decision-making processes connect architecture choices to measured risk, cost, performance, and governance requirements. Migration Center surfaces dependencies and modernization fit, while Cloud Monitoring, BigQuery, and Cloud Billing turn operations and spend into KPIs and dashboards. Teams reduce risk by starting with a non-critical but typical workload with few dependencies to create a blueprint for later waves.
Billing data flows from resources into invoices and BigQuery exports, is structured by reports and FinOps Hub, is enriched with labels and hierarchy, and is then used to forecast and optimize. Metrics are raw data points like CPU or latency, while a KPI measures overall performance calculated from those metrics. Availability tradeoffs follow math: chained availability multiplies tier values, so four 99.9% tiers yield about 99.6%, while parallel availability uses 1 minus the product of failure rates, so two 99% services in parallel yield 99.99%. Responsibility shifts by model: Compute Engine leaves most security with the customer, App Engine and GKE share controls, and Google Workspace leaves most with Google.
Customer success management is the practice of proving that Google Cloud delivers business value and stays healthy through adoption and governance. Cloud Monitoring collects real-time metrics, BigQuery stores and analyzes them over time, and dashboards expose bottlenecks or missed targets. Data flows from workloads to Cloud Monitoring and Cloud Logging, then to BigQuery and dashboards for trends and decisions.
Adoption tracking combines resource signals such as CPU and response time, experience signals such as latency and error rate, business signals such as revenue, and delivery signals such as deployment frequency and time to restore. Health review follows Measure, Analyze, Optimize, and Measure again, comparing service level indicators (SLIs) against SLOs. Without governance, resources and costs accumulate without control, so naming, labeling, access, and cost standards are enforced with labels, tags, and security marks. Proofs of concept validate scope before broad rollout, and rightsizing with autoscaling, committed use discounts (CUDs), Spot VMs, and lifecycle management converts findings into lower cost.
Cost optimization is the shift from fixed Capital Expenditure (CapEx) for owned hardware to variable Operational Expenditure (OpEx) for pay-as-you-go use. Every virtual machine, container, dataset, and network flow creates usage data that flows into Cloud Billing, where labels, budgets, quotas, and recommendations connect spend to owners. Traditional CapEx overprovisions for peaks and pays for idle capacity, while OpEx scales with use but still bills network egress by utilization.
| Control | What it governs | How it enforces |
|---|---|---|
| Cloud Billing budgets and alerts | Actual versus planned spend | Sends notifications; does not block use |
| Quotas | Quantity of shared resources | Blocks provisioning beyond the limit |
| Organization Policy Service | How resources may be used | Restricts products or configurations |
Right-sizing adjusts Compute Engine machine types from Cloud Monitoring signals and Recommender suggestions. Discount choice follows predictability: resource-based CUDs commit to capacity for one or three years, sustained use discounts apply automatically, and Spot VMs offer more than 80% lower prices for fault-tolerant work but move from running to preempted when capacity is reclaimed. Cloud Storage lifecycle rules move aging data to lower-cost tiers, and single-region buckets cost less than dual- or multi-region buckets.
Exam tip: Cloud Billing budgets and alerts only notify stakeholders and are subject to billing-data delay; quotas and Organization Policy constraints actually block or restrict provisioning.
Business continuity is the design of data protection and recovery so work resumes within agreed time and loss limits. RTO defines acceptable downtime, while RPO defines acceptable data loss, with tighter targets requiring more redundancy. Objects in dual-region and multi-region buckets are stored in at least two separate areas, with Turbo replication giving 15-minute RPO in a dual-region.
Firestore scheduled backups run daily or weekly with up to 14 weeks retention and restore to a new database, while point-in-time recovery (PITR) reads up to seven days back at one-minute granularity for surgical repair. Backup and DR Service backs up Persistent Disk snapshots to Cloud Storage, and safeguards such as object holds, soft delete, Object Versioning, and retention locks protect against deletion. Standby choice follows RTO: cold standby keeps only backups with lowest cost, warm standby keeps a minimal replica, and hot standby runs full active-active environments behind a global load balancer with Cloud Spanner, Cloud SQL high availability, and managed instance groups for near-zero RTO at highest cost.
Continue with this study guide: Analyzing and defining business processes
Blue-green deployments on Google Kubernetes Engine switch all traffic from the old server set to the new server set in one switch after verification and temporarily require up to twice the production servers with backward-compatible database schemas, while rolling deployments update servers in place and shift traffic gradually from old to new servers. Choose blue-green when avoiding interruption during cutover outweighs the cost of duplicate capacity, and choose rolling when the goal is early detection with a smaller capacity footprint.
Cloud Billing budgets and alerts only send notifications about actual versus planned spend and do not block use, while quotas block provisioning beyond the limit and Organization Policy Service restricts how resources may be used. Use budgets and alerts for visibility into spend, quotas to limit quantity of shared resources, and Organization Policy Service to restrict products or configurations.
Connectivity Tests in Network Intelligence Center analyzes the configured path from a source endpoint to a destination endpoint without sending live traffic, while Virtual Private Cloud Flow Logs records actual sampled traffic entering and leaving virtual machines per subnet. Use Connectivity Tests to check whether traffic can travel across Virtual Private Cloud routes, peering connections, and virtual private network tunnels, and use Flow Logs with Firewall Rules Logging to see actual traffic and which rule allowed or denied it.
Cloud Build compiles and tests code, Artifact Registry stores versioned container images, and Cloud Deploy promotes the same image digest through development, staging, and production without rebuilding it. The commit triggers the build, Cloud Build creates the image and pushes it to Artifact Registry and runs unit tests and Container Structure Tests, then Artifact Analysis scans for vulnerabilities and Binary Authorization signs the image with an attestation after approval, while configuration and secrets are injected at deploy time from Secret Manager.
Last updated