pollen-alert

Pollen forecast notification via SILAM and Pushover.

Queries the SILAM THREDDS API for pollen concentration at a given location, classifies the risk level using EAN thresholds, and sends a push notification to iPhone/Apple Watch.

How it works

1. Fetches the pollen forecast for each configured allergen from SILAM (regional northern Europe dataset, ~2.5 km resolution; falls back to pan-Europe dataset, ~10 km)
2. Finds the peak concentration in grains/m³ within the lookahead window
3. Sends a Pushover notification if the peak exceeds the configured threshold

For BIRCH, risk is classified using EAN (European Aeroallergen Network) symptom-onset thresholds (doi:10.1007/s40629-025-00357-5):

Level	Grains/m³	Pushover priority
Very High	≥ 200	High (1)
High	≥ 69	High (1)
Moderate	≥ 23	Normal (0)
Low	≥ 7	Low (-1)
Very Low	≥ 2	Low (-1)

Dataset resolution is cached so each run makes exactly one HTTP request after the first.

Configuration

User config is in pollen-alert.toml. Specify location either by city name (geocoded via Nominatim) or by explicit coordinates — if both are present, coordinates take precedence:

# Option A: city name (geocoded on first run, result cached)
city = "Your City"

# Option B: explicit coordinates (geocoding skipped)
# latitude = "YOUR_LATITUDE"
# longitude = "YOUR_LONGITUDE"

lookahead_hours = 24
fetch_interval_hours = 1
notify_hours = [5]

[[allergens]]
name = "BIRCH"
threshold = 81

Secrets and deployment env vars:

Variable	Default	Description
PUSHOVER_TOKEN	—	Pushover app API token (required)
PUSHOVER_USER_KEY	—	Pushover user key (required)
CACHE_FILE	/cache/dataset.json	Path for dataset resolution cache
READINGS_FILE	/cache/readings.json	Path for readings cache
GEOCODE_CACHE_FILE	/cache/geocode.json	Path for geocoding cache
METRICS_PORT	—	Port for Prometheus metrics (enables server mode)
CONFIG_FILE	/config/pollen-alert.toml	Path to config file

Allergens

The SILAM dataset currently provides forecasts for: ALDER, BIRCH, HAZEL, GRASS, RAGWEED, MUGWORT

Configure multiple allergens via POLLEN_ALLERGENS:

POLLEN_ALLERGENS=BIRCH:81,ALDER:50

To list available allergens directly from the SILAM API:

docker run --rm -e PUSHOVER_TOKEN=x -e PUSHOVER_USER_KEY=y ghcr.io/im0rtality/pollen-alert \
  python pollen_alert.py --list-allergens

Local testing

cp .env.example .env
# fill in PUSHOVER_TOKEN and PUSHOVER_USER_KEY

make run

Cache persists in ./cache/ between runs.

Deployment (homelab Kubernetes)

Docker images are published to ghcr.io/im0rtality/pollen-alert on every tagged release. Images are tagged with the full semver version (1.2.3), minor (1.2), and latest (most recent tag).

The Helm chart is published as an OCI artifact to oci://ghcr.io/im0rtality/charts/pollen-alert on every tagged release.

helm install pollen-alert oci://ghcr.io/im0rtality/charts/pollen-alert \
  --set secrets.pushoverToken=xxx \
  --set secrets.pushoverUserKey=yyy \
  --set config.city="Your City"

Or with a local copy of the chart:

helm install pollen-alert ./helm \
  --set secrets.pushoverToken=xxx \
  --set secrets.pushoverUserKey=yyy \
  --set config.city="Your City"

To use explicit coordinates instead of a city name:

helm install pollen-alert ./helm \
  --set secrets.pushoverToken=xxx \
  --set secrets.pushoverUserKey=yyy \
  --set config.latitude=YOUR_LATITUDE \
  --set config.longitude=YOUR_LONGITUDE

To monitor multiple allergens:

helm install pollen-alert ./helm \
  --set secrets.pushoverToken=xxx \
  --set secrets.pushoverUserKey=yyy \
  --set config.city="Your City" \
  --set 'config.allergens[0].name=BIRCH' --set 'config.allergens[0].threshold=81' \
  --set 'config.allergens[1].name=ALDER' --set 'config.allergens[1].threshold=50'

A PersistentVolumeClaim (1Mi) is created automatically to hold the dataset cache across CronJob runs. The job runs daily at 05:00 UTC (08:00 EEST / 07:00 EET).

To trigger a manual run:

kubectl create job --from=cronjob/pollen-alert pollen-alert-test
kubectl logs -f job/pollen-alert-test