SWITCHING CHECKLIST, read 10 Nov 2026
48 tasks, from a year out to switch-off
Download the checklistSwitch guides
This checklist is the sequence a program owner follows from the day a vendor notice arrives to the day the old platform is switched off. It has six phases and 48 items. Each item gives one reason, and a source where a public page supports it. Items with no source come from program practice, not from a vendor document.
The timing assumes a 12 month plan counted back from the date that binds you. For a Cisco Vulnerability Management customer that is the earlier of your contract term end and the Last Date of Support, 30 Jun 2028, in bulletin EOL15812. For a team already inside 12 months, the phases overlap; the guides on this site say which items to shorten first.
Fix the dates and the owner (12 months before the date that binds you)
- Download and date-stamp the vendor notice (for Kenna, bulletin EOL15812). The page is marked Amended and can change again. (source)
- Write down the six printed milestones, from the 10 Dec 2025 announcement to the 30 Jun 2028 last date of support. Your plan counts back from these dates. (source)
- Find the term end date on your own order and compare it with the Last Date of Support. Your contract term may end before the bulletin's final date.
- Match the part numbers on your order to the families in the bulletin. A part number outside the bulletin may be covered by a different notice. (source)
- Check the notice period for non-renewal in your contract. Missing it can renew a subscription you plan to leave.
- Name one program owner and one executive sponsor. Migrations stall when no one owns the date.
- Ask the vendor in writing what happens to API access and data after the term ends. The answer sets how late you can export.
- Open a decision record with the URL, the read date and each question and answer. An auditor or a new director will ask for it.
Export and inventory (11 to 10 months before)
- Request a full export of the asset, vulnerability, finding and fix models in jsonl. The export API accepts json, jsonl or xml, and jsonl is the default. (source)
- Ask for closed findings explicitly. The default export returns open items on active assets only. (source)
- Download each file within 30 days of the request. Search IDs expire 30 days after the request. (source)
- Record row counts against the console and hash every file. A file that downloads can still be incomplete.
- Ask support in writing for the daily full export of assets and vulnerabilities. The documentation says it can be arranged through tech support. (source)
- Export the connector list, asset groups and logged score overrides. They document your sources and the judgments that shaped the ranking. (source)
- List every scanner, ticketing system and report that depends on the platform. Each dependency needs a home on the new side.
- Store the raw files where a leaver cannot delete them. They are the evidence after the subscription ends.
Write the policy and choose (9 to 7 months before)
- Take 12 months of closed critical and high findings and record why each was urgent. The counts show the true weighting of the old program.
- Write the ranking policy as plain sentences, one rule per sentence. Sentences survive a platform change; weights do not.
- Record the old score scale (Kenna: an integer from 0 to 100). No published mapping exists to the scales of other platforms. (source)
- Write the SLA tiers and the exception rules as they are today. You need them to test the vendors.
- Shortlist two to four platforms against your policy, not against a feature list. A short list keeps the pilots affordable.
- Ask every vendor for the documentation page behind each claim. Marketing pages and documentation are different evidence.
- Ask for pricing in writing and note which vendors quote only. Pricing rarely appears on public pages.
- Run a pilot with your own exported data and your own policy sentences. A demo on sample data proves little.
Build (6 to 4 months before)
- Connect the three largest sources first. They carry most of the volume and most of the risk.
- Load all open findings, assets, owners and first seen dates. The queue must be right from day one.
- Load live exceptions with approver, date and expiry. An exception without these fields fails an audit.
- Load 12 months of closed findings and archive the rest. Trends need history; the archive needs no platform.
- Rebuild each policy sentence as a rule and name the rule after it. Named rules can be explained to an auditor.
- Replay the 12 months of closed findings through the new rules. It shows whether urgent items land in the top tier.
- Map the SLA clocks to the original first seen dates. A restarted clock hides overdue work.
- Build the ticket mapping from old ticket IDs to CVE and asset. Re-linking beats closing and reopening.
Run in parallel (4 to 2 months before)
- Name the system of record for each week. Two systems opening tickets create duplicate work.
- Turn off duplicate ticket creation on the platform that is not the record. Engineers stop trusting a queue that duplicates itself.
- Compare the top 50 findings by owner every Friday. A fixed sample makes differences visible.
- Tag every difference with one cause. Only unknown causes should block cutover.
- Write the exit rule before the run starts. Without one the run does not end.
- Switch ticket re-linking one engineering team at a time. A mistake then affects one team.
- Test four exception cases: active, expired, expiring and rule-level. They are where migrations lose records.
- Tell the report owner which platform each number comes from. Numbers will move; readers need the reason.
Cut over and wind down (2 months before to 3 months after)
- Run the final full export from the old platform on the cutover date. It is the last complete snapshot. (source)
- Compare open findings, severity counts and live exceptions on that date. Matching totals show nothing was dropped.
- Disable, do not delete, the old integration accounts. The accounts may be needed to explain old tickets.
- Revoke old API tokens once the final export is verified. Unused tokens are an exposure.
- Report the old and new count for the same date. A step change explained at the time is accepted.
- Archive the raw exports, the mapping file and the verification page with hashes. The audit period may outlast the subscription.
- Name an owner and a retention date for the archive. An archive with no owner gets deleted.
- Hold a review 90 days after cutover and update this checklist. The next migration will start from your notes.
Cite this
Switching Checklist, After Kenna, read 10 Nov 2026. https://bestvulnerabilitymanagementtools.com/checklist/
Sources
- Download and date-stamp the vendor notice (for Kenna, bulletin EOL15812)., read 10 Nov 2026
- Request a full export of the asset, vulnerability, finding and fix models in jsonl., read 10 Nov 2026
- Ask support in writing for the daily full export of assets and vulnerabilities., read 10 Nov 2026
- Export the connector list, asset groups and logged score overrides., read 10 Nov 2026
- Record the old score scale (Kenna: an integer from 0 to 100)., read 10 Nov 2026
Last reviewed 10 Nov 2026.