The integrations don’t solve it, they implement the solution that core provides.
The integration should mark an entity as unavailable if it lost connection. It is up to the integration to implement that correctly. If the integration doesn’t do that, the only option at your disposal is to look at the timestamp properties of the sensor and make a judgement call. There is nothing HA core can do about this.
Presumably you are using the InfluxDB integration, and therefore this is an InfluxDB problem. How does the integration decide when to push a new data point to the InfluxDB file? I haven’t looked at the code but I expect it looks for the state_changed event, which is only fired when the state changes from one number to a different number. If the behavior should be different and you want it to push a data point whenever the entity has a value reported (and not necessarily changed) then the integration needs to be updated to instead look at the state_reported event. Or if you want it to push a new data point every 30 seconds (for example) that’s also the integration’s responsibility.