Reliability Labs

Database Connection Contention and Recovery Lab

Model bounded connection pools, lock waits, retries, replica lag, and failover ambiguity in a deterministic browser-only timeline.

Local workbench

Input stays in this browserRun with Ctrl/⌘ + Enter

Deterministic browser-only database model

Database Connection Contention and Recovery Lab

Explore how bounded connection pools, lock waits, retries, replica lag, and failover interruptions interact—without connecting to a database.

This is not a benchmark, capacity calculator, database dry-run, failover exercise, RPO/RTO calculator, or production verification. Steps, requests, and delays are teaching units rather than measured production behavior.

Six fixed teaching scenarios

Choose a production pressure pattern

Selecting a scenario loads its inputs and runs the same deterministic model.

01 Connection and workload

Model the admission boundary and how long each transaction retains a connection.

Pool sizeAvailable database connections · 1–100 Incoming requests per stepNew modeled requests · 0–200 Transaction durationModeled steps while work is active · 1–20

02 Lock waits and retries

Introduce deterministic contention, wait timeouts, and a finite retry budget.

Lock contentionRequests selected for modeled contention · 0–100% Lock hold durationModeled lock-hold steps · 0–20 Lock wait timeoutSteps before a modeled timeout · 1–50 Retry limitMaximum retries for one request · 0–5

03 Replica and failover

Add delayed replica visibility or a bounded interruption to inspect recovery risks.

Replica readsModeled reads routed to a replica · 0–100% Replica lagVisibility delay in modeled steps · 0–50 Failover starts at stepLeave blank for no interruption Failover pauseModeled interruption steps · 0–20 Simulation lengthTotal modeled steps · 1–120

Ctrl/⌘ + Enter to run

Healthy bounded load is ready. Nothing leaves this browser.

Deterministic result

Connection pressure and recovery timeline

Within the selected teaching boundary

Step	Arrived	Active	Lock wait	Queued	Timeout	Retry	Rejected	Committed	Replica visible	Ambiguous	Failover	Risk

PostgreSQL 18

Evidence to inspect next

MySQL 8.4 LTS

Evidence to inspect next

Copyable educational summary

How to use it

Choose one of six deterministic scenarios or enter a bounded teaching configuration. The same input always returns the same timeline and engine-evidence prompts.

Read the lab examples and evidence guide, then return to Production Database Incident Troubleshooting.

This is not a benchmark, capacity calculator, database dry-run, failover exercise, RPO/RTO calculator, or production verification. Nothing leaves this browser.

Related tools

Keep the workflow moving

These local-first tools often pair well with Database Connection Contention and Recovery Lab.

Available tool

Java Concurrency Budget Lab

Model fixed pools, virtual threads, resource guards, deadlines, and cooperative cancellation with a deterministic browser-only timeline.

Open tool

Available tool

Kubernetes Rollout Capacity Lab

Model readiness, surge, unavailable replicas, drain time, rollout stalls, and version capacity with a deterministic browser-only timeline.

Open tool

Available tool

Message Delivery Semantics Lab

Simulate ACK timing, crashes, bounded retries, redelivery, duplicate effects, and stable idempotency entirely in your browser.

Open tool

FAQ

Database Connection Contention and Recovery Lab questions

Does this lab connect to a real database?

No. It is a pure deterministic browser model with no network, database, credential, storage, remote execution, or telemetry access.

Can the lab size a production connection pool?

No. Counts and steps are explicit teaching inputs, not workload measurements, benchmarks, forecasts, or capacity recommendations.

Does the recovery step calculate RTO?

No. It marks a transition in the teaching model and is not elapsed time, an availability result, a failover exercise, RPO, or RTO.