I've got a device which, due to a firmware bug, sometimes does this:
As you can probably imagine from the fact that the values are in watts, this not not healthy for said device, and I'd like to get an alert when it happens. I've figured out a way of doing it but I thought I'd ask for feedback on whether it's the way to do it. The idea was to use a Threshold helper to check when it goes outside bounds (upper = 5000, lower = 4000), and then connect it to a Statistics helper with max_age = 60 minutes and trigger an automation when count > 3 or 4, something like:
binary_sensor:
- platform: threshold
name: "Grid Power Excursion"
unique_id: uniqueid__grid_power_excursion
entity_id: sensor.grid_active_power
lower: 4000
upper: 5000
sensor:
- platform: statistics
name: "Grid Power Excursion Count"
unique_id: uniqueid__grid_power_excursion_count
entity_id: binary_sensor.grid_power_excursion
state_characteristic: count_on
max_age:
hours: 1
automation notify_grid_power_excursion:
- id: uniqueid__notify_grid_power_excursion
alias: Notify of excessive grid power excursions
triggers:
- trigger: numeric_state
entity_id:
- sensor.grid_power_excursion_count
above: 3
actions:
- action: notify.notify
data:
title: Grid Power Excursion
message: Grid power use is oscillating +5kW / - 4kW
Is that a sensible way to do it, or have I missed an integration/helper/utility that'd do it better?
An update on this after a few days testing, the above doesn't actually work because it triggers continuously when the power usage is outside the given range, not just once when it crosses the threshold. Unfortunately there doesn't seem to be a sensor type that deals with this, requiring an automation for each range-exceeded event that increments a counter, which can then be examined by a statistics sensor:
counter:
grid_power_excursion_counter:
id: uniqueid__grid_power_excursion_counter
name: "Grid Power Excursion Counter"
initial: 0
step: 1
automation check_grid_power_excursion_above:
- id: uniqueid__check_grid_power_excursion_above
alias: "Check for grid power excursion above maximum"
triggers:
- trigger: numeric_state
entity_id: sensor.grid_active_power
above: 5000
for: "00:01:00"
actions:
- action: counter.increment
target:
entity_id: counter.grid_power_excursion_counter
automation check_grid_power_excursion_below:
- id: uniqueid__check_grid_power_excursion_below
alias: "Check for grid power excursion below minimum"
triggers:
- trigger: numeric_state
entity_id: sensor.grid_active_power
below: -4000
for: "00:01:00"
actions:
- action: counter.increment
target:
entity_id: counter.grid_power_excursion_counter
sensor:
- platform: statistics
name: "Grid Power Excursion Count"
unique_id: uniqueid__grid_power_excursion_count
entity_id: counter.grid_power_excursion_counter
state_characteristic: count
max_age:
hours: 1
So this increments the counter each time the power threshold is crossed, and the statistics sensor records how many of these threshold-exceeded counts have occurred in the last hour. A final automation (not shown) then acts on the events-per-hour going over a certain level.
This works, but if there's a less Rube-Goldberg way to go about it I'd be interested in hearing about it.
