SpaceX is making this look easy. Given the frequency and cadence of launches by SpaceX, it's just a matter of time before they suffer a failure during boost phase. I think it will be interesting to see the difference in how SpaceX handles a failure in an established program versus what NASA does.
NASA usually shuts down further launches of vehicles that suffer a problem and studies things for months or even years. I'm betting SpaceX will postpone further launches for a couple or three weeks, find the problem and fix it if necessary, then get back on the launch schedule.
At this point, Falcon 9 Block 5 might be the most fired and most reliable rocket on the planet. It has surpassed Atlas V in that regard, but I'm not familiar enough with Soyuz' various evolutions and forms to say it has surpassed that platform as well.
It is the lowest cost platform ever to launch either unmanned payload or manned vehicles. It has a sufficient cadence and backlog that in the event of a failure it is fairly easy to swap a lower priority/criticality payload onto the next flight in order to prove that a fix is viable to solve the previous problem.
Add to that the fact that they have had the opportunity to examine landed boosters, something never before done with any other rocket, and inspect both engines and superstructure for wear and fatigue, and it is easy to see why the Block 5 iteration is the most reliable rocket in world history.