Murphy's Law

Social
1

I noticed this morning that my energy costs were a lot higher than usual. Unfortunately I had to go out for the day and could not investigate.

After arriving back home I discovered that the automation that changes my energy tariff from peak to off-peak failed to run yesterday evening. Why?

Because I restarted home assistant at a few seconds before 9pm - the time my tariff changes to off-peak for the whole weekend.
:man_facepalming:

It was relatively easy to re-calibrate my grid import energy utility meters that updated my daily costs back to normal.

I also copied all the 5 minute LTS from the peak sensor to a spreadsheet ready to set them to 0 and update the off-peak sensor. However actually making the change causes the statistic editor to go back to the current time. I realised that to make the hundreds of changes I need to would take far too long.

I might come back to it in a week when it has been down sampled from 5 minute data to hourly data. Less than 24 changes that way.

What a bonehead move.

Ever done anything similar?

2

Lucky for you it was just data. I had a reboot that skipped turning my 600W dehumidifier off. I noticed the next day. That was actual money. The good news: I learned from HA statistics that the 300W setting the dehumidifier worked just as well and I also realized that I should put a safeguard in the automation.

3

Good point about making automations more tolerant to this. I’ve done things like this in automations which I want to run once a day, but only once:

  trigger:
  - platform: time
    at: 01:00:00
  - platform: time
    at: 02:00:00
  - platform: time
    at: '14:00:00'
  - platform: time
    at: '22:00:00'
  condition:
  - condition: template
    value_template: '{{ now().strftime(''%x'') != as_timestamp(states.automation.[this automation name].attributes.last_triggered)
      | timestamp_custom(''%D'') }}'
4

Unfortunately that won’t work in this particular case. Peak to off-peak and vice-versa happen at strictly defined times & days.

5

Good point; my example isn’t exactly right for this use case. But I was thinking maybe shorten the times to just a few minutes apart, to cover brief outages like an update. You would lose a little data during that time HA was down anyway, but at least once things come back up the tariff change would be applied.

6

It wouldn’t be easy. This is the trigger:

  trigger:
  - platform: state
    entity_id: binary_sensor.peak_rate
    not_to:
      - unknown
      - unavailable

As it triggers on a binary template sensor changing to on or off.

- name: "Peak Rate"
  unique_id: de98f85a-61be-46e7-92fd-d34320b80e4e
  icon: "mdi:power-plug"
  state: >
    {% if states('binary_sensor.is_dst')|bool(0) %}
      {{ (now().weekday() < 5) and ( (7 < now().hour < 11) or (16 < now().hour < 22) ) }}
    {% else %}
      {{ (now().weekday() < 5) and ( (6 < now().hour < 10) or (15 < now().hour < 21) ) }}
    {% endif %}