Decay heat arises from the full range of decay of fission product and activation product. There are betas, alphas, gammas, and even a few delayed neutrons.
As you move to through the periodic table to higher numbers of protons, you find you need a larger number of neutrons to make it stable. Helium is 2 protons and 2 neutrons. Lead is 82 protons and 122 to 126 neutrons, depending on the isotope. When a nucleus fissions, typically the results have too many neutrons to be stable. This means they will be radioactive. Betas and alphas are ways to adjust this to come to a stable isotope. But they can also emit neutrons, and they often emit a couple gammas on the way.
The fission products build up during operation. They don't go away at shutdown, but continue to decay.
At the moment of a reactor shutdown, depending on the design, decay heat is typical a few percent of full power, typically in the range of 4% to 6%. After a few minutes it will have fallen to 1% to 2%. The decay curve is very complicated because it is the result of many different isotopes with very different half lives. The specific design of the reactor, and the operating history prior to shutdown, have strong effects on the curve. So it's not possible to give an exact value for it. Probably the delayed neutrons are the fastest to decay, with half lives mostly less than about 2 minutes. Though a few hang in there with half lives of a couple hours. After that the gammas are probably the fastest decaying, though again, a few hang in there with intermediate half lives.
So the exact fraction of the decay heat each decay mode produces is changing over time, and is different for different reactors, and different for different operating histories.
Some reactor designs are such that they can passively dissipate the decay heat after a shutdown. Usually this is only for quite small reactors, in the few megawatt range or less. For example, pool type reactors are often designed this way. If you have a reactor about 1 meter on a side in a pool with 100 tons of water, you can probably count on the decay heat being absorbed by the pool. At least for many days which should give you time to correct whatever the problem is. Many research reactors are designed this way.
Typically a large power reactor, say above 1000 MW thermal, will have too much fuel for this to work. At 2% decay heat that's still 20 MW. It means the heat is enough, and the heat dissipation poor enough, that temperatures can rise drastically. You can get melting of metal components. You can get production of hydrogen with the possibility of explosion. You can even get the metal components catching on fire and sucking the oxygen out of the cooling water. This can raise the temperature in the core to extreme levels, say in the few-thousand degrees range.
Some reactor designs combine the coolant and moderator. The fuel is in a big pressurized tank in which you pump in water and get out much hotter water, or possibly steam. It's not possible to separate coolant from moderator in such designs.
Some molten salt reactor designs include the ability to dump the entire working material of the core into a "catcher" under the core. This is designed with a geometry that splits the material into sub-critical regions, and passively absorbs the decay heat for arbitrarily long periods. This is where the burst valve you mention comes in. A plug of material melts if the reactor ever goes above its operating temperature, and the molten fuel drains out. Nobody has built anything with such a design, but it seems on paper to be a good idea.
The rest of this is about CANDU reactors. You can get lots of other info on them here.
https://canteach.candu.org/Pages/Welcome.aspx
In a CANDU reactor, the coolant and moderator are separate. The coolant is in pressurized horizontal fuel channels. The fuel is loaded into the channels from the end of the reactor. The moderator is in a non-pressurized tank so that the moderator fills the space between the fuel channels. The fuel channels are typically pressurized to about 9 MPa or so, and are at temperatures just below boiling, approximately 250°C. The moderator is typically about 60°C. There is typically about 50 tonnes of moderator.
In early designs of CANDU there was a shutdown by dumping the moderator. There were huge valves at the bottom, with a catch tank. The moderator could be drained in some small time like 1.5 seconds, basically a little more than the time it takes to fall 6 meters. However, accident analysis convinced people it was a bad idea. Dumping 50 tonnes of cold water just at the moment an accident was happening turned out to be sub-optimal.
So CANDU has two fast acting shutdown systems. They have control rods, typical absorber rods (maybe Gadolinium, Boron, or Cadmium) that are dropped into the core to absorb neutrons. And they have a liquid absorber injection system that injects neutron absorbing material (typically Gadolinium or Boron) into the moderator.
Also, CANDU has an emergency coolant injection system (ECIS). This is a large tank of cold water that stands by to replenish the coolant in the event of a loss of coolant accident. And, in a severe situation, there is a fitting that permits connecting a fire hose to the reactor to supply coolant.
