Controlling 3800W 230V Water Heater Power for Stable Temperature—Microcontroller vs Analog

I have a resistance (~3800 watts) to boil water. The idea is to control the power supplied to regulate the output water temperature. The voltage supplied is 230v.

What is the best way to regulate power in order to get always the same water temperature?
To set up the regulation point should I use a microcontroller or analogue electronics? I don't have much experience in power electronics.
Any help will be much appreciated. Thanks.

NOTE: 230V CAN KILL YOU. MAKE SURE YOU GET SOMEONE TO VERIFY YOUR CIRCUIT IS SAFE. IF YOU CAN, USE AN ISOLATION TRANSFORMER.
The first thing you have to decide is how you are going to measure the temperature. You have many options- thermistors, diodes, RTDs, thermocouples, semiconductor sensors. Depending on your choice you could easily get away without a microcontroller. You may find an IC intended for the application. Heck you could get away with a thermostat.Here is a circuit that uses just a 555.Another choice you have to make is how you want to control the temperature. There is a bang-bang approach  or you could try some form of proportional control(extending all the way to PID.

The only thing I can add to Aubrey's usual excellent advice is that unless you use something simple like a thermostat (which is what most domestic water heaters still use, but are fairly imprecise and have a fairly large "dead-band") you should use something that uses Zero Voltage Switching.  This switches the power at the zero point on the AC waveform so you don't get high-current transitions which cause RF interference.  Otherwise you will need to put in extra suppression components (usually series inductors and parallel capacitance) to minimise the interference.And I would reiterate Aubrey's first paragraph of safety advice.  It sounds like you have not done much of this sort of thing before.  Take great care and get someone qualified to help you if you can.

Theoretically, boiling water will stay at 100degC (or whatever its nominal boiling point is) until has boiled away, so do you need a controller at all?

Thank you all for your advice.
I still have to research what you have suggested, but do you mean I can manage such power with the 555 circuit proposed? I have read about using a triac optocoupler zero-cross, what do you think?What do you suggest in terms of safety? It is the most important...
Apologise for the delay

Jose...- Triac optocoupler zero cross - give us a link to what you read and we can comment- As Aubrey pointed out if you want to boil water, you don't need a control, if you want water at a certain controlled temperature then you will need a controller of some sort- 3800 W at 230V  = 16.5 Amps.  You will need something that can switch this sort of current, be it a relay or a triac- Optocouplers are great for keeping the control part of your circuit isolated from the actual power side.-  555 circuit  - it looks good, but note that all parts of the circuit will be at mains voltage, so your thermistor temperature sensor will have to be isolated from the water, which may not be easy.  Also note that it specifies a 2A load max. - to run your 16A load will require a much beefier triac which may need more gate current than a 555 can supply.  You could maybe get your 555 to drive a relay to switch the heater element, which would give you isolation as well.- Safety - make sure all metal parts are grounded, insulate live parts as much as possible, make sure no live parts can be touched through holes, make sure no water can reach live parts.  I reiterate -if you're not experienced at working with mains voltages, get someone qualified to help you.  In some countries (Australia where I am being one of them) it is a legal requirement.

Zero crossing means that the triac will start to conduct as the ac voltage crosses zero. This is an attempt to reduce radiated noise since no current is flowing at those times. However zero crossing means that you are working with a bang-bang approach. You cannot adjust the current using phase control.
For either approach there are a series of triac drivers that I have always found to be very effective. For non-zero crossing look at the MOC301x/302x range. For zero-crossing look at the MOC306x range. There are many options in the MOC30xx range. Note that all of them are triac drivers. There will always be an external  triac. They also have the advantage of providing isolation.And please heed David’s warnings. As I said in my first post- 220VAC can kill.

 Great, now I understand it better. Thank you so much,
About safety, I will check the design with an expert before testing.

220VAC (or any other mains) can also damage your instruments like an oscilloscope. They must be rated to work with the high voltage. Also keep in mind that the negative of your ‘scope is connected to  earth/ground through the power connection.

One more thing. If you go with a triac, keep in mind that there is a volt drop, say 1V(check the data sheet).  At 16 A the triac will be dissipating 16W- it’s going to get HOT even at 1/4  that (even in the unlikely event that there is only a 0.25V drop). You should plan for a heat sink.