The 300-615 DCIT doesn't test whether you can configure a Nexus switch or stand up an ACI tenant. It tests whether you can look at a broken environment, some show command output, or a fault code, and figure out what's actually wrong. That's a different skill than the one most CCNP study guides train for, and it's why candidates who breeze through Official DCCOR 300-615 DCIT sometimes stall here. This page walks through what the current v1.2 blueprint actually covers and how to practice for a diagnostic exam rather than a configuration one.
DCIT is built around five domains, and Cisco publishes the weighting directly:
Network (25%) — troubleshooting OSPFv2/v3, MP-BGP, PIM, and FHRP; switching issues in RSTP+, LACP, and vPC; VXLAN EVPN overlay problems; and a substantial ACI section covering fabric discovery, access policies, VMM domain integration, tenant policies, packet forwarding, and external connectivity.
Compute Platforms (25%) — UCS rack and blade troubleshooting, including chassis/power/IOM infrastructure, VLAN and pool/policy issues, SAN connectivity and FC zoning inside UCS, packet flow from server to fabric, VIC adapter and firmware interoperability, and firmware upgrade failures.
Storage Network (15%) — Fibre Channel physical layer issues (switched fabric initialization, buffer credit starvation, NPV/NPIV, VSANs), plus FC services like FCID, zoning, device alias, and Cisco Fabric Services.
Automation (15%) — troubleshooting EEM and Scheduler, Bash/guest shell on NX-OS, REST API calls, JSON/XML encoding issues, and problems surfacing through Python, Ansible, Terraform CLI, or Intersight-driven automation.
Management and Operations (20%) — firmware and interoperability issues, Nexus Dashboard and Intersight integration, network security (fabric binding, port security, AAA/RBAC, first-hop security, MACsec), ACI security domains and role mapping, and compute/storage security.
Notice that Network and Compute Platforms carry the same weight and together make up half the exam. If you're strong on NX-OS but weak on UCS Manager and UCS blade infrastructure, that gap will cost you more than most candidates expect.
This isn't a first certification for anyone. Candidates sitting DCIT have almost always already cleared DCCOR (350-601), the CCNP Data Center core exam, and are now completing the concentration requirement. Most are working data center or network engineers who spend time in Nexus 9000/7000 environments, UCS domains, or MDS SAN fabrics — this exam rewards people who've actually had a fabric go sideways at 2am, not people who memorized a Cisco Press chapter. A smaller group takes it specifically for the standalone Cisco Certified Specialist – Data Center Operations credential, without necessarily finishing the full CCNP.
DCIT is a 90-minute exam, and Cisco doesn't publish an exact question count or passing score — both can shift between deliveries. Candidate reports consistently put it in the 60-70 question range, mixing multiple-choice, multiple-answer, drag-and-drop, and fill-in-the-blank items. What trips people up isn't the format, it's the density: many questions present a chunk of show command output, a topology diagram, or an ACI fault, and ask you to identify the root cause or the next diagnostic step — not just "what does this command do." If your practice has only been flashcard-style command recall, you'll be underprepared for the amount of output-reading this exam demands.Strengthen your preparation by taking a Systems Security Certified Practitioner practice exam.
Don't just drill for the right answer — drill for the reasoning. When you get a question about, say, a vPC inconsistency or an FC buffer credit issue, work out why the symptom points to that cause before checking the explanation. Time yourself in blocks that mirror the pressure: at roughly 60-70 questions in 90 minutes, you have a little over a minute per question, but the output-heavy items eat more of that budget, so you need buffer time from the easier automation or management questions. Track your domain-by-domain accuracy rather than your overall score — a 90% average that hides a 60% on Storage Network is a much bigger problem than the aggregate number suggests.
The most frequent one is treating the Automation domain as a coding test. It isn't — it's troubleshooting automation failures, so you need to recognize what a malformed JSON payload or a broken REST call looks like in output, not write Ansible playbooks from scratch. The second common gap is underestimating the ACI subsections under Network; candidates comfortable with traditional NX-OS routing and switching often haven't spent equivalent time on APIC fault codes, VMM domain issues, or ACI external connectivity, and that shows up hard on exam day. The third is neglecting UCS blade-specific troubleshooting (IOM, fabric interconnects, server pools and boot policies) in favor of general compute concepts — DCIT wants UCS-specific diagnostic instincts, not generic virtualization knowledge.
Build a personal "symptom to cause" table as you study — one column for the observable symptom (a specific show output, an ACI fault, a UCS alert), one for the root cause, one for the fix. This exam is fundamentally asking you to map symptoms to causes fast, so studying in that format matches how the exam actually questions you. Also, don't skip the security subsections under Management and Operations (5.3-5.6) just because they feel like an afterthought tacked onto a 20% domain — AAA/RBAC and fabric binding issues show up as standalone questions, not just footnotes.
Work through the practice questions below by domain rather than in random order, starting with Network and Compute Platforms since they carry the most weight, then check your Storage Network and Automation accuracy separately before your exam date — those are the domains most candidates under-prepare for relative to their weight.
Last updated on Oct, 9 2026
