OutSystems Platform Ops Engineer flashcards

34 cards. Read the term and the definition here. The deck at the top of this page is optional.

LifeTime console

Central management console for all environments in a stage set. It registers environments, builds deployment plans, provisions users, and provides cross-environment monitoring and audit of deployments.

OutSystems stage

A grouping of environments such as Development, Quality, or Production. Deployments move forward through stages in order, and each stage can contain more than one environment.

Environment isolation

Each environment has its own Platform Server and database catalog. An application deployed to one environment is not visible in another until it is deployed there too.

Deployment Controller role

Service inside each environment that receives deployment packages and applies them. If it is stopped or unreachable, deployments fail even when the application server is healthy.

Solution pack

Portable file containing an application's modules and their references, used to move work between environments. It is uploaded to the target by the deployment process.

Platform Server version parity

All environments in a stage set should run the same Platform Server version. Version skew between LifeTime and an environment blocks deployments and complicates troubleshooting.

Deployment plan

LifeTime's diff-based selection of applications and module tags to move from source to target. It resolves dependencies and flags conflicts before anything is applied.

Module tag

Immutable snapshot of a module used as the unit of deployment. Deployments always reference tags, never mutable development revisions, which makes releases reproducible.

1-Click Publish

Service Studio action that compiles and deploys directly to the connected environment. It bypasses LifeTime, so it is normally restricted to non-production environments.

Rollback by redeploy

The platform does not roll back automatically. To revert, deploy the previous tag of the affected modules as a new deployment and verify the result.

Dependency conflict

Occurs when a referenced producer module or extension differs between source and target. The plan highlights producers that must be deployed first to resolve it.

Destructive schema change

Entity model synchronization applies schema changes during deployment, but dropping columns or tables can lose data and needs explicit review before the deployment is confirmed.

Stage skipping

Deploying from Development straight to Production without passing Quality. It requires an explicit LifeTime setting and is recorded in the audit trail.

Post-deployment compile

Final deployment step that compiles and warms up the application on the target. Skipping or failing it leaves first requests slow or the application unavailable.

Platform Server tiers

Standard installation separates the database server, the Platform Server nodes running Application Server and Deployment Controller, and the load balancer in front of server farms.

Server farm shared cache

Farm nodes must share database and file storage for the compiled application cache. Without sharing, each node compiles independently and first requests after deploy are slow.

Service Center logs

Per-environment log viewer filtering by module, error type, and time. It is the primary source for runtime exceptions and integration failures in that environment.

Integration logs

Records of REST and SOAP calls with payloads and duration, enabled per integration. They grow quickly, so retention must be pruned deliberately.

Performance Monitor

Service Center view showing screen and integration timings per module. It separates slow queries from slow client rendering before infrastructure changes are made.

Controller unavailable error

Message indicating the Deployment Controller service is down or unreachable. It is an infrastructure fault, not an application defect, and blocks all deployments to that environment.

Sticky sessions

Load balancer setting that routes a user to the same node, or central state storage that removes the need. Without either, users lose session state between farm nodes.

Environment health checklist

Recurring checks of Platform Server version, free disk, database connectivity, Deployment Controller status, and last successful deployment timestamp for each environment.

LifeTime roles

Administrator, Developer, and IT Manager assignments managed centrally and separate from per-environment roles. A LifeTime administrator can manage environments without app-level developer rights.

Service Center roles

Per-environment roles including Administrator, Developer, and Operator. They govern who can deploy, configure, and operate within that specific environment.

SAML 2.0 authentication

Federated login option configured per environment. Changing the provider invalidates existing sessions, so cutover should be scheduled outside business hours.

Built-in authentication

Platform-managed user accounts stored in the environment database. Common in Development, while production typically federates to Active Directory, SAML, or OpenID Connect.

Per-environment secrets

Connection strings and API keys stored in environment configuration rather than in modules. LifeTime environment settings are the supported location for these values.

Deployment audit trail

LifeTime record of who deployed what and when, combined with Service Center role-change logs. Together they provide evidence of controlled change management.

Least privilege deployment

Developers hold Developer role only in Dev and Quality. Production deployments run under an operator or pipeline account, and Administrator is limited to platform operators.

Scale out versus up

Scale out adds Platform Server nodes behind the load balancer; scale up increases CPU and RAM on existing nodes. The database usually saturates first, so tune it before adding nodes.

Aggregate column trimming

Aggregates fetch only attributes referenced downstream. Adding unused attributes widens queries and increases memory, so trim each aggregate to the fields actually used.

Capacity planning inputs

Peak concurrent users, requests per user per minute, average and p95 response time, database CPU, and app-node CPU. Targets keep p95 within SLA with failover headroom.

Synchronous integration cost

A synchronous REST call inside a screen blocks the request thread until it returns. Making it asynchronous where the result is not immediately needed frees capacity.

Bounded cache expiration

Reference data cached with an explicit expiration rather than indefinitely. Unbounded caches grow until they exhaust memory and behave like a leak.