From software problem to a clear next step.
Choose a topic and what you need to do next. Add a search phrase to find a specific problem.
Browse 20 topics
Explore the playbooks
1500 guides · Page 36 of 60
-
Triage cancellation races at the job commit boundary
Determine whether cancellation won before or after a job commit by comparing request time, transaction or write evidence, acknowledgement, readback, and visible terminal state.
-
Prioritize post-commit cancellation by reconciliation risk
Rank cancellation races by unknown durable effects, duplicate-prone retries, downstream status decisions, and the availability of a trustworthy commit readback.
-
Investigate cancellation before and after a durable commit
Build a reproducible cancellation brief by pausing a job before commit, after commit, and during status acknowledgement while tracing records and terminal mapping.
-
Verify cancellation before, after, and during job commit
Check cancellation before commit, after commit, and with incomplete readback, confirming terminal states, durable effects, retry eligibility, and downstream interpretation.
-
Define committed, canceled, and unknown job outcomes
Preserve cancellation ordering, commit ownership, readback requirements, terminal states, retry rules, and boundary fixtures for jobs where a stop request can race durable work.
-
Triage job claim delay before measuring runtime
Locate ready-job claim delay by comparing enqueue, ready, poll, lock contention, claim, start, worker availability, and execution timestamps.
-
Prioritize ready-job claim delay by queue-age impact
Rank claim delay work by time spent waiting before execution, blocked downstream actions, queue reach, worker capacity, lock contention, and evidence of avoidable dispatch delay.
-
Investigate job claim delay with quiet and busy controls
Build a claim-delay brief by matching jobs across worker availability, poll cadence, queue routing, lock acquisition, claim, and start.
-
Verify ready jobs claim promptly across worker conditions
Check claim timing with quiet workers, busy workers, lock contention, and a normal queue path, confirming ready-to-claim thresholds and runtime separation.
-
Track ready-to-claim age as its own job measure
Preserve event definitions, claim-age guards, worker and lock context, matched dispatch fixtures, and recurrence checks that keep pre-execution wait separate from runtime.
-
Triage shutdown interruptions by checkpoint and recovery owner
Locate what a worker shutdown left unresolved by comparing signal, drain state, active phase, checkpoint, process exit, requeue, restart, and terminal outcome.
-
Prioritize shutdown interruption by lost-work and replay risk
Rank worker shutdown work by unowned jobs, duplicate side effects, checkpoint recovery, dependent blockage, and confidence that restart semantics are safe.
-
Investigate worker shutdown at safe and mid-step boundaries
Build a reproducible shutdown brief by stopping a worker before a checkpoint, after a commit, and during a known active step, then tracing restart ownership.
-
Verify worker shutdown before and during active job steps
Check graceful drain, safe-point stop, mid-step interruption, and restart recovery, confirming checkpoints, ownership, terminal states, and no lost or repeated effects.
-
Document worker shutdown, drain, and restart ownership
Preserve shutdown modes, safe checkpoints, active-step recovery, claim ownership, requeue rules, and fixtures that detect lost or repeated jobs after worker exit.
-
Triage dependency waits by prerequisite state and release edge
Locate a job dependency wait failure by comparing prerequisite IDs, edge mapping, terminal states, release events, wait age, timeout, and dependent claims.
-
Prioritize dependency waits by blocked graph consequence
Rank dependency wait work by dependents blocked, jobs released before prerequisites finish, graph reach, wait age, and confidence in prerequisite state and edge mapping.
-
Investigate dependency release across success and failure paths
Build a reproducible dependency brief by replaying prerequisite success, failure, cancellation, missing, and delayed terminal states through the release edge.
-
Verify dependency waits for every prerequisite outcome
Check all-success, failure, cancellation, missing, and delayed prerequisite cases, confirming dependent wait, release, timeout, terminal state, and downstream eligibility.
-
Document prerequisite states and dependency release rules
Preserve graph revisions, edge and fan-in semantics, prerequisite outcome mapping, wait-age guards, timeout behavior, and fixtures for dependent jobs.
-
Triage batch-size regression with matched records and resources
Locate a batch-size regression by comparing fixed input coverage, query behavior, memory, runtime, worker concurrency, errors, and completion counts at two sizes.
-
Prioritize batch-size regression by correctness and resource risk
Rank batch-size work by skipped or repeated records, memory or connection pressure, worker blockage, recovery cost, and confidence that size caused the observed change.
-
Investigate batch-size changes with a bounded size sweep
Build a reproducible batch-size brief by testing previous, current, and larger sizes against fixed ordered records while tracing selection, query, resource, and commit behavior.
-
Verify batch sizes for coverage, runtime, and resource headroom
Check lower, current, and higher batch sizes with fixed records, confirming exact item coverage, query and checkpoint behavior, terminal outcomes, and resource guards.
-
Document a tested batch-size range and its resource guard
Preserve size limits, input ordering assumptions, coverage fixtures, query and resource measures, checkpoint behavior, and review triggers for future batch changes.