# Bring open findings, exceptions and 12 months of closed findings, and archive the rest

PROGRAM GUIDE, read 10 Nov 2026

**The short answer.** Loading every record from the old platform slows the new one and imports old errors. Bring all open findings, all live exceptions, the SLA start dates and 12 months of closed findings for trend lines. Archive the raw export, with a hash, for the audit period you must cover.

## Three tiers of history

Sort the old data into three tiers. The first tier is what the program needs to run tomorrow: open findings, assets, owners, live exceptions and the first seen date of every finding. Load all of it. The second tier is what the program needs to show progress: closed findings with their closed dates for the last 12 months, which give you a trend line and a mean time to remediate that does not start at zero. Load these as closed records.

The third tier is everything else: older closed findings, retired assets and superseded scores. Archive it as raw files. The old export files are the evidence, and a new platform's interpretation of them is not.

## Why not load everything

Older records bring older errors: assets that were renamed, scanners that were retired and duplicate findings that were never merged. They also make the first deduplication run slow and hard to check. If your policy or your auditor requires a longer window, load it, but load it into a separate view or dataset.

Keep the original score out of the new ranking. Cisco's score was an integer from 0 to 100, and your new platform ranks on its own scale. Store the old score as a read-only column labeled with its source and date, so that trend charts do not mix two scales.

## Get the files out

For Cisco, request the asset, vulnerability, finding and fix models through the Request Data Export API. The defaults return open items on active assets, so ask for closed findings explicitly. Exports arrive in gzip, in json, jsonl or xml, and the search ID is valid for 30 days, so download the files within that window. The export guide covers the steps.

For other sources, check the export routes before you commit. Brinqa's ledger entry lists a REST BQL API that works asynchronously, BigQuery views through BrinqaDL (which its marketing page calls an add-on with separate pricing) and file connectors for CSV, JSON and XML input. A CSV export from the Brinqa platform UI: none found on 2026-11-10 in the pages read.

| Tier | Contents | Where it goes |
| --- | --- | --- |
| 1: run | Open findings, assets, owners, live exceptions, first seen dates | Loaded as live records |
| 2: show progress | Closed findings, last 12 months | Loaded as closed records |
| 3: evidence | Everything older; raw exports; old scores | Archived files with hashes |

## Decide with the people who will ask

Ask three people what they will want to see in two years: the auditor, the security director and an engineering lead. The auditor wants decisions with dates. The director wants a trend. The engineer wants to know why an old ticket exists. Each answer maps to a tier above.

Write the decision down with the retention period and who owns the archive. An archive nobody owns gets deleted when the subscription ends.

- Load all open findings and live exceptions.
- Load 12 months of closed findings for trends.
- Archive the rest as raw files with hashes.
- Name an owner for the archive and a retention date.

## Measuring what you carried

After the load, check three numbers against the old platform on the cutover date: the count of open findings, the count of open findings by severity tier, and the count of live exceptions. They should match, with differences you can explain. Add a fourth check on closed history: the count of findings closed in each of the last twelve months should be the same on both sides.

Then compare a trend you report today, such as mean time to remediate for the critical tier. The number will move a little, because deduplication and asset matching differ. Report both numbers once, with the cutover date, and note the cause. Readers accept a step change that is explained at the time. They do not accept one that is discovered later.

Keep the verification file with the archive. It is a one page record that says what was loaded, when, from which export files, and how the totals matched. When someone asks in 2028 why the trend line has a break in 2027, that page answers in a minute.

Costs are a factor too. Some platforms price by asset or by finding volume, and some charge for data retention or for add-on data layers, as Brinqa's marketing page says of BrinqaDL. Ask each vendor whether loading twelve months of closed findings changes the price, and whether closed records count toward any limit. The answer can change which tier of history you choose.

Whatever you decide, test a load of the closed history on a sample before the full run, and time it. A load that takes a weekend is fine. One that takes three weeks needs a plan.

## Sources

- Request Data Export, Kenna Platform API: https://apidocs.kennasecurity.com/reference/request-data-export.md (read 10 Nov 2026)
- Downloading Export Files: https://apidocs.kennasecurity.com/docs/downloading-export-files.md (read 10 Nov 2026)
- Brinqa API: https://docs.brinqa.com/docs/brinqa-api (read 10 Nov 2026)
- Brinqa data layer (vendor page): https://www.brinqa.com/platform/data-layer (read 10 Nov 2026)

Last reviewed 10 Nov 2026. Canonical: https://bestvulnerabilitymanagementtools.com/guides/choosing-what-history-to-bring/
