Remove Personal Information from KNIME Server Log Files

Masks IPs, email addresses, and usernames in KNIME Server Logs for all standard cases.

How it's used
Most conveniently, it's executed as a Data App on KNIME Server where the logs are stored and the workflow will fetch logs automatically.
You can also execute the workflow locally and enter the server credentials manually to access the current server logs.
In case the server is not reachable or you have a set of logs already, you can execute the workflow locally and provide the logs as a zip file.
Please remember that this is a workflow you can modify and extend to your specific needs.

What it does
With this Data App, KNIME Server log files can be depersonalized in a way such that IPs, email addresses, and usernames are masked. Emails and IP addresses (IPv4 and IPv6) are both detected via regex and any match will be masked (see linked websites below). Usernames are determined from
- KNIME Server internal DB
- Foldernames in the /Users folder
- Usernames that occur in the localhost access logs.
There can be no guarantee that this workflow removes "all personal information" , since an organization may have naming protocols for different aspects of KNIME that KNIME can know nothing about. The list of search/replace criteria can therefore be extended by providing a CSV File containing a column "username" to extend the list.

Usernames and IP addresses found in the localhost_access logs will receive a mapping throughout all logs (Eva Musterfrau -> , Max Mustermann -> , ... and 127.0.0.2 -> , 127.0.0.3 -> , ...), everything else that matches will be masked with a generic tag ([e-mail] -> , [e-mail] -> , ...).

Possible processed log files
- localhost_access
- localhost
- catalina
- the executor log (KNIME log) can also be processed but is very large and not usually needed for support purposes, therefore processing this log is by default NOT done unless the setting is overwritten.

External resources

• Email Regex
• IPv4 and IPv6 Regex