Kinda surprised you guys aren't talking about the Asplodey Kaboom that happened this weekend, with the SpaceX Crew Dragon capsule.
This exact capsule was up at the ISS a few weeks back, and splashed down in the Atlantic after this test launch and docking (unmanned). SpaceX retrieved it and were initially very pleased with the lack of penetration of salt water on the design. So much so that Musk tweeted about it. In their pursuit of reuse of as many components as possible, they put this recovered prototype capsule on a test stand and fired up the various hypergolic engines on the craft. There are Draco engines which provide orbital maneuvering and fine tuning during docking procedures, and there are SuperDraco engines which are the launch escape system for the craft in the event of a rocket malfunction, and they were intended to also serve as propulsive landing engines rather than parachutes.
We still don't have the full story, but the capsule was completely destroyed during one of the engine ignitions. Don't know if it was the Dracos, or the SuperDracos.
But both use hypergolic fuel, and both use the culprits from the previous SpaceX rocket explosions: COPV's. Carbon Overwrap Pressure Vessels. The most recent rocket explosion (AMOS-6) is thought to be due to fine gaps in the COPV mesh allowing liquid oxygen to creep inside the gaps between the carbon mesh and the aluminum pressure bottle inside it, and then any thermal stress/shift/flex results in a very energetic reaction with that oxygen. Boom. But the COPV's in the Crew Dragon aren't sitting in a bath of liquid O2.
This exact capsule was docked to the ISS less than 2 months ago, and its Draco and SuperDraco systems were fueled and operational. That same kaboom could have happened a few weeks earlier, killing all ISS inhabitants and possibly causing irreparable harm to the station (if NASA and Roskosmos even had the drive to rebuild it in the face of such a loss).
I'm hoping that this was a controlled edge-case test with unexpected results; they had noticed that the SuperDraco fuel lines would ice up during certain phases of its mission, perhaps they were testing the results of this edge case and got a result outside of what they were expecting, in a catastrophic way, but it would still be a known failure mode they were just confirming.