Java Database Connectivity

It isn't just a passive storage locker where it writes a log entry when a flow finishes.

Kestra uses the relational database as an active message queue broker via JDBC polling.

Wait, using a database as a message queue? I thought that was what things like RabbitMQ or Redis were for.

How does a database act as a queue?

Well, I mean, on the, on the, the paper cut one, it...

From what I read, it exploited a, like the JDBC driver, so the authentication mechanism, and completely bypassed it.

So as long as you could hit the server, you could then- ... essentially bypass the mech, the, the authentication, get access, or, or an RCE so you have route, or you'll have, you have access to the actual server, which then attackers were using that to pivot within the environments.

And they had detected, what was it, Huntress? I think Huntress had detected that there were, like, already couple people already breached from this.

If you're building pipelines, you can now create, store, and manage them in Git folders in Preview, while regular folder storage for pipelines has reached general availability.

For developer tooling, the Rust SDK for BigQuery is in Preview, and there's an updated Simba JDBC driver available.

Also, a quick bug fix, Google has restored support for configuring daily token quotas for BigQuery generative AI functions.

Let's check in on Looker and Looker Studio.

It's kind of a middle ground.

You can just throw in your SQL statement and it takes away all the pain that JDBC API can be in terms of exception handling, in terms of not having name parameters and the like.

So there's always kind of a gray area to pick one.

Yeah.

What do you work on there at the moment? I started from the outside, object mapping, spring data integration.

For the last two years, it was great work digging out decades-old technology again.

You and I both have a history of doing stuff with JDBC, Java Database Connectivity.

Mostly, it's dealt with in terms of relational databases, but There's a lot of references to SQL and SQL CLI in there, but nothing in JDBC claims you can only use this with relational databases.

the following day.

Early indicators of compromise include alerts involving PC-app.exe, post-exploitation activity, missing or truncated server.log files, and specific JDBC or database utils error strings in your logs.

Paper cut print servers have a grim precedent.

Prior paper cut flaws were abused by CLOP and lock bit affiliates and a compromised application server is a ready made foothold for lateral movement and ransomware because it sits at the intersection of user networks and sensitive document workflows.

So a fresh zero day should be handled as an assumed compromised event for internet facing instances.

What should you do if you're running Papercut NG or MF? Apply the emergency patch for version 25, especially version 25 release 2 immediately and restrict the application server's web interface to trusted internal IPs via firewall or network ACLs if it was internet-facing.

Hunt the published IOCs for pc-app.exe execution, missing or truncated server.log files, and those JDBC database utility errors in your log files.

Rotate any of the credentials you have with access to the papercut server, even if no IOCs are found, and treat any internet-facing instance you have as potentially compromised, pending a clean hunt.

An authenticated attacker could craft malicious JDBC connection parameters to achieve remote code execution in the connector container.

Please check your deployments, refer to the security bulletin for the JDBC driver vulnerability, and ensure you are using the updated drivers to avoid privilege escalation.

Now on to workflow orchestration.

As part of its branding evolution, Cloud Composer is now officially known as the managed service for Apache Airflow.

And you mentioned MCP.

I was just doing some research for another client who actually are the folks that built the ODBC and JDBC drivers 20-odd years ago.

So they know a thing or two about data access and provisioning and so forth.

A little bit.

And as you may know, they own Simba Technologies.

And Simba, for those who love the history of our industry, Simba wrote the ODBC and JDBC drivers that are still the workhorses of data management all over creation.

Drivers are important.

They come in handy for being able to optimize things.