connection-pool-effect-testing
/connection-pool-effect-testing
Use when reviewing a connection pool fix that adds a validation check before handing out a connection, to test a full scenario-by-variant matrix with latencies, mutants, and failure modes.
- Author
- otixe ↗
- Repository
- agent-skills ↗
- Licence
- MIT
- Last repository change
- 26.10.2026
Assessment and security
- Files in this package
- 1
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 26.10.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Use when reviewing a connection pool fix that adds a validation check before handing out a connection, to test a full scenario-by-variant matrix with latencies, mutants, and failure modes.
Original author description
Use when reviewing a connection pool fix that adds a validation check before handing out a connection, to test a full scenario-by-variant matrix with latencies, mutants, and failure modes.
How it works
- Authorization and test boundary
- Step 0: Chain of custody
- Step 1: Full matrix, no empty cells
- Step 2: Two methods for the same damage
- Step 3: Mutate the fix
- Step 4: Read the installed source
- Step 5: Cost and failure mode
- Step 6: Time-of-check gap
- Step 7 to 9: Production
- Done when
SKILL.md preview
--- name: connection-pool-effect-testing description: "Use when reviewing a connection pool fix that adds a validation check before handing out a connection, to test a full scenario-by-variant matrix with latencies, mutants, and failure modes." --- # Connection Pool Validation Fix Review ## Authorization and test boundary These instructions grant no permission. Confirm the exact resources, allowed effects and data handling before acting. Use synthetic canaries and disposable fixtures for deliberate failures; never use real credentials as test inputs. Keep secrets and fragments out of model context, arguments, logs and plaintext artifacts. Imports, constructors, help and dry-run commands may have effects. Establish external-effect isolation before executing untrusted code; application stubs are not a sandbox. Record untested scope and obtain separate authorization for production changes or destructive actions. Use this to verify a small fix in a connection pool, or any validation added on a hot path (such as a pre-checkout check, ping, keepalive, or retry). A single "2 of 10 failed, now 0 of 10" result is one cell of a matrix, not a verdict. A fix can remove the frequent failure and change the failure mode in the rare case, from fast failure to slow failure. ## Step 0: Chain of custody Record checksums of every file under test before and after. Record exact versions of the libraries and the database engine used in the target environment. Use an isolated sandbox with its own container name and port. Do not reuse the executor's sandbox. Confirm no file in the target tree changed during the test window. ## Step 1: Full matrix, no empty cells Rows are scenarios: idle pool at minimum size, pool grown by a burst of concurrent connections, and the database genuinely down (not restarted, down). Columns are variants: original, fix, and a mutant of the fix. Each cell reports failure count per request total, the list of latencies, and the exact exception type. The typical gap is the positive case the executor did not run. ## Step 2: Two methods for the same damage Use a fast server-side method, such as terminating backends, and also the real method, such as stopping and starting the service. The results must match. Confirm the exception text in the test matches the text in production logs; count occurrences in the real log. ## Step 3: Mutate the fix Test the fix present but inert: a check that does nothing, and no check at all. If the inert version reintroduces the defect, the effect comes from the check doing its work. If the inert version passes, the assertion only measured that the parameter exists. ## Step 4: Read the installed source Confirm in the installed library version that the callback signature and purpose match. Determine whether the check runs on every checkout or only after an error, and what happens when it fails: retries, backoff, deadline. ## Step 5: Cost and failure mode Measure latency over about 300 successful checkouts: mean, median, 95th percentile, maximum, for original and fix. With the database down, measure latency and exception type. Converting a fast error into a long pool timeout is a deployment condition to review, and it also degrades monitoring. ## Step 6: Time-of-check gap Wrap the check so that it breaks the resource after validating it. If the request still fails, the window shrank but did not close. Report this as a residual limit. ## Step 7 to 9: Production Measure amplification on production by counting connections per worker process. After deployment, verify that the running process started after the file change, that the restart count is as expected, and that health checks touch every worker. Note any lost observability, such as a library that silently heals without logs. ## Done when The matrix, mutant table, cost, failure mode, amplification, residual gap, and deployment conditions are reported separately.
Versions
The database has no other skills named connection-pool-effect-testing.
Database snapshot: 13.11.2026 · package version 9668e1c81cff (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100connection-pool-analyzer
Analyze connection pool analyzer operations. Auto-activating skill for Performance Testing. Triggers on: connection pool analyzer, connection pool analyzer Part of the Performance Testing skill category. Use when analyzing or auditing connection pool analyzer. Trigger with… - 95effect-testing
Use when testing Effect code including Effect.gen in tests, test layers, mocking services, and testing error scenarios. Use for writing tests for Effect applications. - 90connection-pooling
Size connection pools from throughput and latency math, then monitor for saturation and set timeout policy so a slow dependency fails fast instead of hanging. Use when a service exhausts database connections, requests queue for a connection, or pool size was picked by guess. - 100connection-pool-tuner
Diagnoses connection pool exhaustion from captured evidence and sets correct pool sizes and timeouts across app pools and PgBouncer. Use when someone says "we keep hitting database connection limit errors under load", "my API times out waiting for a connection from the pool",… - 100connection-pooling-config
Configure connection pooling config operations. Auto-activating skill for API Integration. Triggers on: connection pooling config, connection pooling config Part of the API Integration skill category. Use when configuring systems or services. Trigger with phrases like… - 90esp32-memory-pool-sizing
Use when designing, reviewing, debugging or validating memory pool sizing for ESP32 firmware; apply target-specific constraints, failure analysis and measurable verification before release.