`miso_exporter`

This is a Prometheus/OpenMetrics exporter for MISO data.

Quickstart

Container images are available at Docker Hub and GitHub container registry.

$ docker run -it --rm -p 2023:2023 willglynn/miso_exporter
# or
$ docker run -it --rm -p 2023:2023 ghcr.io/willglynn/miso_exporter
2023/07/08 14:51:12 Starting HTTP server on :2023

Endpoints

GET /lmp returns the Locational Marginal Price (LMP) for each node in each region of the MISO network. There are four kinds of prices:

5m, the realtime 5-minute price
1h, the realtime hourly integrated price
exante, the day-ahead ex ante price
expost, the day-ahead ex post price

There are three components:

LMP, the full location marginal price
MLC, the component of the price due to transmission losses
MCC, the component of the price due to congestion

# HELP miso_price_usd The price for power at a certain place and time
# TYPE miso_price_usd gauge
miso_price_usd{comp="LMP",kind="1h",node="AECI",region="Midwest"} 26.61 1688839200000
miso_price_usd{comp="LMP",kind="1h",node="AECI",region="Midwest"} 26.61 1688839260000
miso_price_usd{comp="LMP",kind="1h",node="AECI",region="Midwest"} 26.61 1688839320000
miso_price_usd{comp="LMP",kind="1h",node="AECI",region="Midwest"} 26.61 1688839380000
miso_price_usd{comp="LMP",kind="1h",node="AECI",region="Midwest"} 26.61 1688839440000
…
miso_price_usd{comp="LMP",kind="5min",node="AECI",region="Midwest"} 46.11 1688845200000
…
miso_price_usd{comp="LMP",kind="exante",node="AECI",region="Midwest"} 32.65 1688842800000
…
miso_price_usd{comp="LMP",kind="expost",node="AECI",region="Midwest"} 32.56 1688842800000
…
miso_price_usd{comp="MCC",kind="1h",node="AECI",region="Midwest"} -0.05 1688839200000
…

GET /load returns the total load, forecast and actual.

# HELP miso_load_total_w The amount of load, forecast or actual
# TYPE miso_load_total_w gauge
miso_load_total_w{kind="actual"} 7.298e+10 1688792400000
miso_load_total_w{kind="actual"} 7.298e+10 1688792460000
miso_load_total_w{kind="actual"} 7.298e+10 1688792520000
miso_load_total_w{kind="actual"} 7.298e+10 1688792580000
miso_load_total_w{kind="actual"} 7.298e+10 1688792640000
miso_load_total_w{kind="actual"} 7.2789e+10 1688792700000
miso_load_total_w{kind="actual"} 7.2789e+10 1688792760000
miso_load_total_w{kind="actual"} 7.2789e+10 1688792820000
…
miso_load_total_w{kind="forecast"} 7.2009e+10 1688792400000
…

GET /renewable_production returns actual and forecast production from solar and wind.

# HELP miso_renewable_production_w The amount of power produced from renewable sources
# TYPE miso_renewable_production_w gauge
miso_renewable_production_w{kind="actual",source="solar"} -5.02e+06 1688792400000
…
miso_renewable_production_w{kind="actual",source="solar"} 1.78845e+09 1688849940000
miso_renewable_production_w{kind="actual",source="wind"} 3.21474e+09 1688792400000
…
miso_renewable_production_w{kind="forecast",source="wind"} 7.952e+09 1688965140000

GET /fuel returns actual production by fuel type.

# HELP miso_fuel_w The amount of power produced using a particular fuel
# TYPE miso_fuel_w gauge
miso_fuel_w{kind="actual",name="Coal"} 3.1159e+10 1688852400000
…
miso_fuel_w{kind="actual",name="Wind"} 4.604e+09 1688852640000

GET /metrics returns all of the above.

Prometheus configuration

A good starting point:

scrape_configs:
  - job_name: 'miso'
    scrape_interval: 2m
    scrape_timeout: 1m
    static_configs:
      - targets:
          - http://localhost:2023
    metric_relabel_configs:
      - if: '{__name__=~"^miso_.*"}'
        action: labeldrop
        regex: "instance|job"

/metrics returns unusually large responses, exceeding e.g. VictoriaMetrics' default maximum scrape size of 16 MB. -promscrape.maxScrapeSize=67108864 seems to work.

As of this writing, miso_price_usd (via /metrics or /lmp) produces 4116 series. This is very high cardinality but it is also very low churn; series are created or destroyed only as the set of nodes changes, and MISO nodes correspond to long-lived capital projects like substations or power plants.