Viessmann vitodens 222 not modulating properly

Regarding the ODR settings, I was quite happy with them, the shift of 8°C and a slope of 1.3 worked pretty well to keep my place at 19/20°C in all conditions.
Though I did a test yesterday evening and dropped the shift by 4°C:

The burner reacted just fine, although this morning it was a bit cold inside so I set it back to 8.

Note that the current outside temperatures are quite low, around 1°C and the burner works like a charm.

I'll wait for the boiler to short cycle again when the temperatures will reach 5/6°C and try the suggestion of @Eastman. What will this test trying to show?

I made a couple of tests over the last few hours. Since the outside temperatures are up these are "ideal" conditions for the boiler to short cycle.

1/ In the below screenshot you see that the boiler was running smoothly for an hour until the power goes up to 30% (yellow arrow) for no apparent reason and then stops. After some analysis, at 8am two things happen, my smart TRV in the bathroom closes and the target indoor temperature goes from 19 et 18°C. I understand that this flow reduction might be the issue. But what I would have expected is for the boiler to simply run at a lower power, from 11% to 9% for instance and not go up and stop eventually.

To confirm this I made a new test at 15.45 (red arrow), the boiler was running smoothly and I intentionally closed 3 TRV, same behaviour.

2/ I tried to leave more TRV fully open. In total there are 11 radiators, all with TRVs.
Before I had

• 2 always fully open (5/5)
• 2 at 4/5
• 2 at 2/5
• 3 are on and off demanding on our usage (two are smart TRVs)
• 2 are always closed

Now I changed to this:

• 4 always fully open (5/5)
• 2 at 2/5
• 3 are on and off demanding on our usage (two are smart TRVs)
• 2 are always closed

Also I added a rule that does this:

When the burner starts, immediately open fully two smart TRVs for 5 min and increment the indoor temp setting by 4°C for 8 minutes.
The goal is to increase the flow when the burner starts, to help him evacuate the starting heat (since the boiler always starts at 60% power) and give him a change to stabilize to a lower power.

This seems to improve the situation but not drastically, here is the graph with this logic implemented:

3/ @Eastman regarding the drop for the DHW storage temp in my initial post, this is simply a shower. I guess that the slow increase folowing the shower is simply the fresh cold and hot water left in the tank mixing over time. The tank is heated only at specifit times, you clearly see when, and at that point it is really fast:

Yes the viessmann api provides a datapoint every minute. A typical burner start hits 60% for a few seconds, but in my graphs you can see that they are not always captured. Despite this, the graphs show some pretty interesting data.
And they always match the data from the app or the boiler display.