Engineering Manager interview questions
100 real questions with model answers and explanations for Senior Engineering Manager candidates.
See a Engineering Manager resume example →Practice with flashcards
Spaced repetition · Hunter Pass
Questions
I would produce a one-page quarterly operating plan with explicit outcomes, capacity, owners, and confidence ranges.
- I would spend days 1-3 sizing only the 12 candidate outcomes with the tech lead and product manager, using S, M, and L ranges rather than estimating all 46 items.
- I would model all 12 engineers at 80% planned capacity, giving 9.6 engineer-equivalents to roadmap work and 2.4 to support, leave, and interrupts based on the prior quarter.
- I would publish three committed outcomes, two stretch outcomes, named decision owners, and a weekly confidence update in the same document.
Why interviewers ask this: The interviewer is evaluating whether the candidate can turn incomplete inputs into a credible quarterly operating contract without false precision.
I would replace the output count with a renewal outcome and make the feature set a tested hypothesis rather than the objective.
- I would set the lagging key result as actual renewal rate rising from 78% to 85% among accounts whose contracts expire in Q4.
- I would track renewal intent as a separate leading metric, targeting a 10 percentage-point rise from a week-1 survey baseline while blocker-related tickets fall 30%.
- I would keep the six features in the initiative list, require 60% weekly adoption for the two target workflows, and review the leading evidence at weeks 4 and 8.
Why interviewers ask this: The interviewer is testing whether the candidate can connect engineering work to measurable customer outcomes while preserving accountability.
I would allocate against quantified risk and publish what will not fit instead of pretending 150 weeks can fit into 120.
- I would reserve 30 weeks for reliability because the risk register links it to the agreed 99.9% availability target.
- I would fund one tooling week first and release the other 14 only if telemetry shows at least 40 team-hours saved per week, keeping payback within about 15 weeks.
- I would provisionally cap feature work at 75 weeks, show the 25 weeks removed, and return the gated tooling capacity if the validation misses its threshold.
Why interviewers ask this: The interviewer is assessing disciplined capacity allocation and the ability to expose opportunity cost with evidence.
I would plan with a 25% interrupt buffer and express the delivery date as a confidence range tied to weekly burn.
- I would classify the prior 26 weeks of unplanned work into support, compliance, and dependency requests so the buffer reflects recurring demand rather than intuition.
- I would place 75% of capacity into the committed plan and give finance P50 and P80 dates, with P80 as the external date.
- I would review buffer consumption every Friday and re-scope within 48 hours if four-week rolling interrupts exceed 25%.
Why interviewers ask this: The interviewer is checking whether the candidate uses historical demand to make commitments resilient rather than hiding uncertainty.
I would commit the billing migration and time-box validation of the experiments rather than book unvalidated upside as capacity.
- I would verify the $400,000 retention exposure with finance and put the 10-week migration on the committed roadmap with one accountable tech lead.
- I would allocate one engineer-week to instrument and test the strongest experiment during weeks 1-2, with a written threshold of at least $250,000 expected annual value.
- I would list the other two experiments as uncommitted options and reopen allocation at the week-3 quarterly checkpoint only if the threshold is met.
Why interviewers ask this: The interviewer is evaluating portfolio judgment under constrained capacity and unequal evidence.
I would commit only the 90% and 70% outcomes and show the others as dated options with explicit evidence gates.
- I would put the first two outcomes in the customer deck with their current confidence, delivery windows, and owners rather than unsupported P80 labels or exact dates.
- I would require a completed prototype by week 3 for the 55% item and a resolved external API dependency by week 4 for the 40% item.
- I would send sales a weekly confidence table and pre-agree that only items at 70% or above enter customer commitments.
Why interviewers ask this: The interviewer is testing whether the candidate protects delivery credibility while giving commercial teams usable information.
I would turn the dependency into a jointly owned milestone with an early contract test and a funded fallback.
- I would agree an API schema and passing consumer-driven contract test by week 2, with one named owner from each team.
- I would reserve 10 engineer-days for an adapter against the current API and release the remaining 14 days only after the week-2 gate passes.
- I would record the dependency, 60% confidence, decision date, and fallback in both quarterly plans and review it twice weekly until week 6.
Why interviewers ask this: The interviewer is assessing whether the candidate manages dependencies through artifacts, gates, and capacity rather than status chasing.
I would run a written replan that funds the legal deadline by removing lower-value scope, not by silently increasing utilization.
- I would ask legal for signed acceptance criteria and split the requirement into must-have and deferrable controls within the first 8 hours.
- I would compare remaining capacity with the 18-week demand, pause the lowest-value outcome, and preserve the 20% interrupt buffer.
- Within 48 hours I would publish the revised outcome table, displaced value, new P80 dates, and approvals from product and legal.
Why interviewers ask this: The interviewer is checking whether the candidate can absorb a fixed constraint transparently without creating hidden overcommitment.
I would anchor the goals in developer adoption and time saved, with reliability as a guardrail.
- I would target weekly adoption from 38% to 60% among the 160 engineers and define an active user as completing one platform workflow.
- I would reduce median build wait from 22 to 12 minutes and calculate reclaimed engineer-hours from CI telemetry.
- I would hold platform availability at 99.9% and require no more than a 2% increase in failed builds while adoption grows.
Why interviewers ask this: The interviewer is evaluating whether the candidate can define measurable outcomes for an enabling team without inventing revenue attribution.
I would lead with the outcome gap and use delivery data to explain decisions, not present 7 of 9 as success.
- I would show baseline, target, and actual for all three customer outcomes beside the nine shipped or unshipped items.
- I would identify which two assumptions failed using adoption data and customer interviews, with one owner and two-week validation action for each.
- I would carry forward only work supported by evidence and record one operating change, such as an outcome checkpoint at week 4, for the next quarter.
Why interviewers ask this: The interviewer is testing whether the candidate distinguishes output from impact and converts a review into a better next-quarter system.
I would create an evidence packet for each engineer and separate observed facts from manager judgment.
- I would pull 12 months of project outcomes, design ownership, review contributions, and customer or partner feedback into a dated evidence table.
- I would ask each engineer and two regular collaborators for examples against the same level rubric, then verify claims against artifacts rather than popularity.
- I would write ratings only after an evidence-gap review with the tech lead and flag any conclusion supported by fewer than two independent examples.
Why interviewers ask this: The interviewer is assessing whether the candidate can produce fair, auditable reviews despite incomplete memory and uneven visibility.
I would compare each case against common level anchors and keep only ratings supported by repeated scope and impact.
- Each slide would show the expected level, two dated examples, measurable outcome, and the candidate's specific contribution rather than team output.
- I would test all five against the same bar used for peer cases and downgrade a proposal if its evidence shows one exceptional event rather than sustained performance.
- I would record calibration decisions and missing evidence in the review system within 24 hours so final narratives match the room.
Why interviewers ask this: The interviewer is testing evidence-based advocacy and willingness to correct local rating inflation.
I would explain that eight weeks cannot manufacture sustained senior-level evidence and plan for the next fair review point without promising promotion.
- I would map the prior 6-12 months to the senior rubric, recording that technical delivery is proven while repeated cross-team decision leadership is not.
- I would give them normal roadmap ownership involving two partner teams, but treat its RFC and outcome as one data point rather than an artificial promotion assignment or popularity vote.
- I would submit only when multiple dated examples show sustained next-level scope and impact; otherwise I would document the remaining gap and continue the plan into the next cycle.
Why interviewers ask this: The interviewer is evaluating honest sponsorship, rubric use, and whether stretch work tests the actual promotion gap.
I would select evidence that shows durable cross-team leverage, not the volume of code or the title already held.
- Artifact one would be an RFC adopted by three teams, with the lead's decisions and a measurable reduction such as 25% lower integration time.
- Artifact two would show a capability they built in others, for example two engineers independently owning follow-up designs within six months.
- Artifact three would be a roadmap or risk decision where their influence changed an outcome, supported by named stakeholder feedback and before-and-after data.
Why interviewers ask this: The interviewer is assessing whether the candidate understands staff-level impact and can construct a concise promotion case.
I would reconcile the ratings dimension by dimension against shared artifacts before either manager argues for an overall label.
- I would ask both managers to score the four dimensions separately and cite dated behaviors from the 10-week project.
- I would distinguish the engineer's direct contribution from the project result using the decision log, review history, and partner feedback.
- I would document the agreed rating, one unresolved evidence gap, and a next-cycle observation plan in the review record before calibration.
Why interviewers ask this: The interviewer is testing whether the candidate can normalize ratings through evidence rather than authority or compromise averaging.
I would run a visibility-bias audit and correct any rating not supported by comparable evidence.
- I would compare ratings by location across the same level, tenure, project scope, and four rubric dimensions rather than use the raw averages alone.
- I would sample two artifacts and two stakeholder inputs per engineer, including asynchronous design and support work that meetings may hide.
- I would present anonymized patterns in calibration and require managers to rewrite or change ratings whose evidence standard differs by location.
Why interviewers ask this: The interviewer is evaluating whether the candidate can detect and remediate proximity bias with a concrete review process.
I would create a 12-week succession scorecard with delegated decisions and an explicit handover date.
- I would define five lead capabilities, including planning, design judgment, delivery risk, stakeholder communication, and developing peers, with observable evidence for each.
- Candidate A would own the next planning draft and candidate B the dependency review, then they would swap shadow roles after four weeks.
- At week 8 I would choose the successor from the scorecard, publish a four-week handover checklist, and name a backup for the next six months.
Why interviewers ask this: The interviewer is testing whether the candidate builds leadership continuity through observed work rather than an informal preference.
I would use a six-month readiness plan with real, reversible leadership assignments and monthly evidence reviews.
- In months 1-2 they would facilitate two design reviews and publish decisions within 24 hours, measured by stakeholder clarity rather than meeting attendance.
- In months 3-4 they would own one roadmap dependency involving product, design, and another engineering team, with the current lead acting only as a reviewer.
- In months 5-6 they would run a planning slice for four engineers, and I would decide readiness against a written scorecard plus feedback from at least five partners.
Why interviewers ask this: The interviewer is assessing whether the candidate develops a lead through authentic scope and evidence instead of title-first promotion.
I would standardize evidence collection before writing and measure both manager time and return rate in the next cycle.
- I would introduce a monthly evidence log with four rubric dimensions and links to artifacts, capped at 10 minutes per report.
- Six weeks before calibration I would run a packet-completeness check and peer-review a 20% sample using one checklist.
- For the next cycle I would target manager effort below 20 hours and returned narratives below 10%, then remove any template field not used in decisions.
Why interviewers ask this: The interviewer is testing process design that reduces review burden without weakening decision quality.
I would give them a 90-day evidence plan with two observable outcomes, not a longer list of generic expectations.
- I would translate the five behaviors into the two current gaps, such as independently reducing a dependency and raising the output of two peers.
- For each gap I would name one 90-day assignment, baseline metric, expected artifact, and weekly source of evidence.
- We would review evidence at days 30, 60, and 90, while stating clearly that completing activities does not automatically determine the next rating.
Why interviewers ask this: The interviewer is assessing whether the candidate turns calibration feedback into specific growth evidence without promising ratings.
Locked questions
- 21
You need to add five engineers to a 10-person team in six months: two senior backend, two middle full-stack, and one SRE; recruiting can support 18 interviews per week. What hiring plan do you publish?
hiring - 22
Your four-month hiring funnel for middle backend engineers shows 400 sourced, 120 recruiter screens, 60 manager screens, 24 onsites, 6 offers, and 3 acceptances. You need four starts in the next 12 weeks. Where do you intervene?
funnelhiring-funnelhiring - 23
Over eight weeks, 72% of candidates pass the technical screen but only 18% pass the four-person onsite; interviewer notes average 40 words and often say "not senior enough." What do you change before next Monday?
- 24
You are designing a structured loop for a middle backend role that must own a service handling 2,000 requests per second within three months; the loop is limited to four 45-minute interviews. What do you include?
design - 25
For a senior frontend role, eight interviewers propose testing 17 competencies across five rounds, but candidate drop-off rises sharply after 3.5 hours. You must finalize the loop in two days. How do you reduce it?
testing - 26
A data-engineer candidate requests extra time and screen-reader-compatible materials two days before a four-round loop; six interviewers worry that accommodations will lower the bar. What do you do?
- 27
12 engineers will join an interview panel for the first time over the next 6 weeks; the hiring target is 8 candidates per week and no trainee may decide alone. What calibration program do you run?
program-managementhiringcalibration - 28
After 30 candidates, one interviewer passes 80% of technical screens while the other four average 45%; all use the same 45-minute prompt. You have five days before the next interview block. What do you do?
- 29
In a 25-minute debrief, four interviewers bring scores of 4, 3, 2, and 2 for a staff candidate; two scorecards were submitted after participants saw each other's ratings. How do you run the decision?
- 30
Your preferred senior engineer candidate has a competing offer of $190,000 base; your approved range tops at $180,000, but you can vary a $20,000 sign-on bonus, equity, and start date within 48 hours. What offer do you construct?
- 31
A candidate you want has two offers and must decide in seven days; their notes say growth and technical ownership matter more than a 5% cash difference. Your team of 12 can offer ownership of a new service, but its roadmap is not approved. How do you close?
roadmapownershiproadmapping - 32
You have budget for one hire this quarter: a senior engineer can raise delivery capacity by an estimated 15%, while a dedicated SRE could cut the team's 12 engineer-hours per week of reliability toil by half. The requisition closes in 10 days. Which role do you open?
capacity-planningreliabilitycapacity - 33
For a 13-engineer team, the last 90 days show one deployment per week, 18-day lead time, 22% change failure rate, and 9-hour restore time; product asks for a delivery improvement plan for next quarter. What do you set?
deployment - 34
Your dashboard reports average lead time of 8 days, but over the last 60 days the median is 4 days and P85 is 21 days across 140 work items. What do you present at the monthly review?
- 35
An 11-engineer team has 31 items in progress, median cycle time has grown from 6 to 11 days over eight weeks, and product resists stopping new starts. What WIP policy do you propose for a four-week trial?
- 36
Across the last six two-week iterations, your 10-engineer team completed 52%, 91%, 64%, 87%, 58%, and 83% of planned work; finance needs a 90-day forecast by Friday. How do you forecast?
iteration - 37
For 85 completed items in the past month, active coding took a median 2 days but total cycle time was 12 days; review wait was 4 days and test wait was 3. What change do you fund for the next six weeks?
- 38
The main CI pipeline for 12 engineers takes 35 minutes at P50 and fails non-deterministically on 8% of runs; you can allocate two engineers for four weeks. What oversight plan do you approve?
ci-cd - 39
A team releases every 10 days even though CI passes in 18 minutes; analysis of 50 changes shows a median wait of 6 days after approval because releases are batched. What policy do you trial for one month?
releasesbatch - 40
Change failure rate rose from 9% to 17% across 120 deployments in eight weeks, while deployment frequency doubled; leadership wants to return to weekly releases. What alternative do you propose for the next month?
releasesdeployment - 41
Your 13-engineer team lists 70 technical-debt items, and product will fund at most 20% of next quarter's 130 engineer-weeks. How do you choose the 26 weeks?
- 42
After cycle time became a target, a 9-engineer team split work into tiny tickets and median fell 40%, but customer lead time stayed at 16 days over two months. How do you repair the dashboard?
- 43
Three engineers submit an 18-page RFC for a service expected to handle 5,000 requests per second; eight reviewers leave 94 comments over 10 days and no decision is recorded. What RFC process do you impose for the next review?
decision-makingconcurrency - 44
Three days before an authentication launch, Security finds a session-revocation gap that needs four weeks to remove; compensating controls reduce but do not eliminate account-takeover risk. Who owns the residual risk?
risk-managementlaunchesauth - 45
At quarterly planning, your team identifies 16 technical risks for a 12-week roadmap, from a deprecated SDK to database capacity; leadership wants a one-page register by Friday. What fields and cadence do you use?
capacity-planningquarterly-planningroadmap - 46
Your 10-engineer roadmap has nine milestones over 14 weeks and depends on security review, a data-team schema, and a vendor sandbox; two dependencies have no dates. How do you make the roadmap executable?
milestonesprocurementplanning - 47
Your team must choose an email API vendor within three weeks for 8 million messages per month, 99.95% availability, EU data residency, and a $6,000 monthly ceiling. How do you evaluate three vendors?
procurementapidecision-making - 48
An observability vendor renewal rises from $48,000 to $72,000 per year; your team of 14 has a $60,000 tool budget and uses only 6 of 15 licensed features. The renewal is due in 30 days. What do you do?
procurementobservability - 49
Your 11-engineer team has a $90,000 annual discretionary budget; halfway through the year, $52,000 is spent and requests total another $55,000 for test devices, training, and a contractor. How do you allocate the remaining $38,000 within two weeks?
- 50
A code-scanning vendor costs $24,000 per year and claims it will replace 10 engineer-hours of manual review each week; procurement needs a recommendation in 15 business days, and your tool budget has $30,000 left. What decision process do you run?
procurementconcurrency - 51
At week 7 of a 12-week roadmap, your 10-engineer team has completed 40% of the committed scope and the critical path is two weeks late. What do you do?
schedulingdependenciesscope-management - 52
Two days after your integration tests passed, a partner deployed an unversioned API change that removed a required field; 8% of requests now fail and launch is in five days. How do you respond?
integrationapideployment - 53
Product wants a committed date for migrating 40 TB of tenant data, but no test exceeds 2 TB and downtime, residency behavior, and rollback time are unknown. What do you commit to?
rollback - 54
You reported 80% confidence in a quarter-end launch, but a newly discovered integration issue drops the realistic probability to 35% with 24 days remaining. What do you communicate?
probabilitycommunicationlaunches - 55
An executive wants a six-week delivery for a project your 12-person team estimates at ten weeks because it supports a $2 million sales target. How do you handle the pressure?
soft-skillsestimation - 56
A strategic customer representing 12% of ARR requests a private product fork with unique authorization and reporting rules; the team estimates 18 engineer-weeks to build it and 10 weeks a year to maintain it. What do you do?
estimationauth - 57
Ten days before launch, testing finds 14 defects, including two that can corrupt data, while Marketing has already bought a $180,000 campaign. How do you make the release decision?
launchesreleasestesting - 58
Your team has rolled over 30% of planned work in each of the last five two-week sprints, yet estimates have not changed. How do you fix planning?
agileestimation - 59
Support escalates 35 cases a month to Engineering; each case changes hands three times on average, and engineers spend six hours reconstructing missing customer context. How do you repair the handoff?
escalation - 60
Forty-eight hours before release, three blockers remain: a 4% payment failure rate, incomplete rollback automation, and one missing analytics event. How do you run go or no-go?
releasescommunicationrollback - 61
Over eight weeks, an engineer has missed 5 of 8 commitments, caused four escaped defects, and needs twice the review time of peers. How do you start active performance management?
performance-managementperformancedefects - 62
After six weeks of coaching, an engineer still meets only 50% of agreed expectations. HR asks you to design a 30-day PIP. What goes into it?
coachingdesignpip - 63
During compensation calibration, you recommend a 7% increase for an engineer whose scope grew, but the committee cuts it to 2% and its worksheet omits six months of acting-lead duties. How do you dispute the decision?
calibrationcompensationscope-management - 64
A senior engineer with five years on your product resigns, leaving a ten-person team with only one other person who understands billing. What do you do in the first two weeks?
- 65
A new senior hire is at day 60, has missed three onboarding milestones, and has alienated two reviewers by dismissing repository conventions. What do you do?
milestonesplanningonboarding - 66
After market bands rise, new middle engineers are hired within 2% of senior engineers who have three years of strong performance; several incumbents raise internal-equity concerns. What do you do?
performance - 67
After promotion packets are submitted, the committee adds organization-wide mentoring as a senior criterion even though it was absent from the published rubric used throughout the cycle. What do you do?
mentoringpromotionperformance-rubric - 68
An engineer disputes a below-expectations rating because they closed 25% more tickets than the team median, while review data shows repeated rework and two missed ownership duties. What do you say?
ownership - 69
A senior engineer has missed two roadmap ownership duties, but a workload audit finds 14 hours a week of unassigned coordination across six teams that is absent from their goals. How do you respond?
roadmapownershiproadmapping - 70
Two of your 11 engineers enter active performance management in the same quarter, putting a committed release at risk by about 20%. How do you manage both people and delivery?
risk-managementreleasesperformance - 71
You must choose tenant-based or region-based partitioning for a 40 TB data platform; residency and cross-tenant analytics conflict, and moving tenants becomes costly after 20% migration. How do you decide?
migrationspartitioningconflict-management - 72
A new billing workflow is scheduled to roll from 10% to 50% of customers next week, but Support has no runbook, case tags, sandbox training, or escalation owner. What do you do?
escalationrunbooks - 73
Design requests a three-week interaction polish pass, but engineering has nine days before an accessibility deadline affecting 18% of users. How do you decide?
estimationdesigna11y - 74
Data reports a 6% conversion lift, but engineering finds the experiment excluded 22% of mobile traffic due to an event bug. Product wants to launch tomorrow. What do you do?
experimentslaunches - 75
Two senior engineers both claim ownership of a shared service, and their conflicting changes have caused three rollbacks in one month. How do you reset ownership?
ownershipconflict-managementrollback - 76
After three production emergencies, the same three reliable engineers absorbed 36 after-hours hours while six teammates received none; another high-risk launch is next week. How do you allocate emergency work fairly?
risk-managementlaunches - 77
In four architecture meetings, two remote engineers spoke for under 5% of the time while three office-based seniors made every decision. How do you make the next decision inclusive without slowing it indefinitely?
architecture - 78
Your tech lead has overridden peer design feedback on three consecutive RFCs, and two senior engineers have stopped commenting. How do you intervene?
feedbackdesign - 79
A PM has promised dates to customers three times without engineering estimates, creating 25% sprint churn. How do you repair the working agreement?
agilepromisesestimation - 80
Product, Design, and Engineering each rank a different issue as the top priority two weeks before launch: activation, usability, and crash rate at 1.8%. How do you make one decision?
activationdesignlaunches - 81
Checkout errors jump from 0.4% to 35% ten minutes after a deployment, affecting about $20,000 in orders per minute. What do you do as the manager?
deployment - 82
An eight-engineer rotation receives 18 actionable pages per week, including six between midnight and 6 a.m., and two engineers want to leave on-call. What do you change?
on-call - 83
A severity-one incident has run for six hours as the incident commander in one time zone reaches the end of the shift; recovery is incomplete and the incoming team lacks context. How do you hand over command?
incident-managementincident-commandincidents - 84
During a postmortem, deploy logs show a change at 14:02, customer errors begin at 14:07, chat reports the change at 14:11, and two engineers remember the order differently. How do you reconstruct the timeline?
postmortemdeploymentincidents - 85
Your team shipped 12 regression defects last quarter, and 7 came from the same legacy pricing module. What reliability action do you take?
pricingdefectsreliability - 86
Your service delivers 99.5% availability against a 99.9% customer commitment, while Product wants all ten engineers on a feature projected to add $500,000 ARR. How do you allocate capacity?
capacity-planningcapacity - 87
Seventy percent of alerts last month required no action, but engineers still acknowledge every alert within five minutes. How do you reduce alert fatigue safely?
alerting - 88
Two senior engineers handle 60% of after-hours incidents in a nine-person team because others lack permissions and confidence. How do you rebalance on-call?
incident-managementon-callincidents - 89
A canary deployment reaches 10% of traffic and raises API failures from 0.3% to 4% within eight minutes, but the feature team asks for 20 more minutes to inspect logs. What do you decide?
deployment-strategiesapideployment - 90
A release cannot complete its regional failover test before launch, but delaying would breach a customer deadline; leaders ask for temporary acceptance of the reliability risk. What memo do you require?
risk-managementlaunchesreleases - 91
Within one product area, 12 engineers are split into two pods, but 65% of work waits on the same three backend specialists. How would you reorganize?
reorganization - 92
A hiring freeze removes two planned roles from your ten-engineer team's quarter, cutting expected capacity by 18% after the roadmap was announced. What do you change?
capacity-planninghiringroadmap - 93
You must reduce your area's operating budget by 15% in 30 days without layoffs, and vendor contracts make up 40% of spend. How do you proceed?
procurement - 94
A critical identity vendor is down for three hours, blocking 28% of logins, and its status page gives no recovery time. What do you do?
procurement - 95
A VP starts ranking individual engineers by deployment frequency, and one engineer responds by splitting changes into 20 tiny deployments per week. How do you correct the DORA metric misuse?
deploymentmonitoring - 96
A four-week experiment replacing code review with mandatory pair programming raises lead time from 2.5 to 4.5 days and team satisfaction falls from 7.8 to 5.9 out of 10. What do you do?
experimentscode-reviewprogram-management - 97
After your team misses two customer dates by 18 and 24 days, Product no longer trusts an eight-week forecast. How do you repair stakeholder trust?
stakeholder-managementstakeholder-trustcommunication - 98
After a reorganization, pulse survey clarity drops from 8.1 to 6.0 out of 10 and 7 of 12 engineers cannot name their service owner. What do you change?
reorganization - 99
Your primary test vendor has two outages in 30 days, blocking releases for 19 total hours, but replacing it would cost six engineer-weeks. What do you decide?
procurementreleases - 100
Your product has 180 feature flags, 47 are past their removal dates, 19 have no owner, and two stale flags combined to cause a production incident. How do you restore the flag lifecycle?
incident-managementincidentsfeature-flags