https://api.weather.gov/ no longer returns API status in JSON #823
-
Hi, Prior to the revamp, https://api.weather.gov/ used to return the API status via JSON in the format of "status":"OK" and respond within ~3 seconds. Now this returns HTML/documentation even when making the request with the accept header set to JSON. I see there's now a health check endpoint at https://api.weather.gov/health which seems to take ~12 seconds to respond. It seems to always return a status of "fail" even when there don't seem to be any sort of failure. To my untrained eye, it seems to treat slightly stale/delayed data warnings as a failure. I can't find any sort of deprecation notice for getting a simple status from https://api.weather.gov via JSON like I'm accustomed to, and I couldn't find documentation for the https://api.weather.gov/health endpoint like what exists for other endpoints. Can the original simple status reporting be restored? Or can the /health endpoint be corrected/documented? To add some context to this, we are a subscriber and do weather monitoring/reporting for large manufacturers. We provide an alerting system so they can make decisions on whether to pause production/assembly lines. When the API is truly experiencing issues, we then notify them that they may not be receiving accurate/timely alerts anymore and to check other sources for information on inclement weather. Thank you for your time and consideration,
|
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 2 replies
-
A 200 response from api.weather.gov is probably most similar to the OK json response returned previously. It seems the alert:latency check from the health endpoint might be useful to you if you are looking for something more robust regarding the alerts ingest side and not just api being up. |
Beta Was this translation helpful? Give feedback.
-
The previous root endpoint was never really meant to be a health check. It never even really ensured that everything was OK — it was just a static response to tell that the application server was running, but upstream services required for operation could still be unavailable. Given its effective meaninglessness, we opted to make it more of a resource for visitors presumably wanting to know the API's capabilities. Any sort of OK response from |
Beta Was this translation helpful? Give feedback.
The previous root endpoint was never really meant to be a health check. It never even really ensured that everything was OK — it was just a static response to tell that the application server was running, but upstream services required for operation could still be unavailable. Given its effective meaninglessness, we opted to make it more of a resource for visitors presumably wanting to know the API's capabilities.
Any sort of OK response from
/
is equivalent to the old check. The/health
endpoint exists for more detailed metrics.