Replies: 73 comments 7 replies
-
We are actively working this issue but still trying to narrow it down. Something that would help us is adding this header to your request:
This will give some additional details in the response headers that we can pass along.
If you can provide those in any follow up here it would be very helpful to us. |
Beta Was this translation helpful? Give feedback.
-
I've updated my cURL request in PHP to have
Right now, I seem to have api.weather.gov responses from 23.200.224.17 and that Akamai cache is working fine (current data).
I'll keep watching for more failure.. seems to happen in the mornings. |
Beta Was this translation helpful? Give feedback.
-
Got a stale one:
|
Beta Was this translation helpful? Give feedback.
-
And another sample using a different gridpoint (RC=500 this time)
|
Beta Was this translation helpful? Give feedback.
-
Awesome, thank you so much. This is a weird one but I think I have an idea what's going on now. The good news is we might have just fixed it accidentally while addressing another issue. I'm trying to get these bad objects purged out of Akamai and hopefully I'm right and they don't return. |
Beta Was this translation helpful? Give feedback.
-
Here's another one: -- this time from 104.123.204.17
|
Beta Was this translation helpful? Give feedback.
-
Was that from earlier today? The date suggests it was. We flushed Akamai at 6/25 0100 UTC and I'm hoping that resolves this. Please let me know if you are still seeing this with requests after that point. |
Beta Was this translation helpful? Give feedback.
-
Ahh.,. you're correct. Thought I'd done a shift-reload to refresh the page, but apparently got the old page from earlier in the day. A fresh fetch shows correctly:
I'm hoping your cache flush will do the trick.. I'll keep checking on it. |
Beta Was this translation helpful? Give feedback.
-
Yeah, thanks for all that, and keep us posted if you see more. The root issue is with an internal service that decodes the forecast grids from the field offices. It seems to be randomly spitting out bad data, which is how you saw a forecast that was last modified in the future. Unfortunately a full fix for this is going to take some time. The last hotfix added some sanity checking to what we pull from this internal service, so you may see 500 errors occasionally as this code does its job, but it will hopefully stop totally bogus data from making it to the edge and getting stuck in Akamai like this. |
Beta Was this translation helpful? Give feedback.
-
I just got a stale page. I've been getting them intermittently for the last week or so. I only just found this issue, so I don't have cache headers, but I will add them and check. |
Beta Was this translation helpful? Give feedback.
-
See below. This was retrieved about 5 minutes before the timestamp on this comment (it's many hours out of date). Once I start getting cached data I can get it pretty reliably, I assume akamai is pinning me to a particular endpoint. Request
Response headers
data
|
Beta Was this translation helpful? Give feedback.
-
@duncanwerner This one does not appear to be a bug. MTR has not actually issued a new forecast grid since 1107Z. |
Beta Was this translation helpful? Give feedback.
-
This is the correct current forecast (which I got by changing headers):
(truncated this time) |
Beta Was this translation helpful? Give feedback.
-
@duncanwerner Oh, sorry, now I understand what you're pointing at. Pretty sure that's a bug. It's correctly recognizing that the forecast grid hasn't changed but glossing over the fact that we've rolled over into the morning period. Something I failed to account for because they usually re-issue grids well before this would be noticed. |
Beta Was this translation helpful? Give feedback.
-
The first time I had this problem I was getting data that was more than 24 hours old. It's possible there was no new forecast in that period, but it seems unlikely. It might be that that was the issue you were discussing at the top of this thread, and that you fixed it. Unfortunately my first attempt to debug it cleared it up, so I don't have any useful records. More recently I have gotten a couple of forecasts like the one I sent, which show an unexpected period 1 in the past. It's not a big deal to work around it, although it is confusing. (I will admit I had been ignoring timestamps and just showing periods in order). API is great, btw, thanks for maintaining it! |
Beta Was this translation helpful? Give feedback.
-
I'm getting either stale results on this endpoint: https://api.weather.gov/gridpoints/PDT/23,41/forecast or the following response: I've noticed that I get updated information locally, but then stale information from my server. Additionally, this link does not have stale results: Detailed info for this gridpoint was current up till yesterday and now it's pulling in data from October 22nd. Can I provide any detailed information to help with this? Is there any workaround? |
Beta Was this translation helpful? Give feedback.
-
I'm having the same issue with location Kearney, MO. The data returned is from Christmas 2020. |
Beta Was this translation helpful? Give feedback.
-
Ditto. I've given up on this API. NWS doesn't seem to be rushing to fix, I'll pay a little extra for an API that supported. |
Beta Was this translation helpful? Give feedback.
-
Your frustration is valid. The small team has been rushing non-stop since we inherited the project, and are about to release a major upgrade that will stabilize the entire platform. After that, we should be able to address these critical issues more quickly. |
Beta Was this translation helpful? Give feedback.
-
These specific issues should be reported as operational outages as described here. |
Beta Was this translation helpful? Give feedback.
-
Thanks @scadergit and @StephenClouse. I appreciate your efforts on this and the fact that this API is available. Let me know if there's anything I can do or test with the upgrade. |
Beta Was this translation helpful? Give feedback.
-
I also appreciate the effort and am glad to help in any testing. |
Beta Was this translation helpful? Give feedback.
-
I sincerely apologize for my earlier post. It was uncalled for, especially being a programmer myself. I do understand issues are being worked on and, I too, appreciate the effort. Wish you guys had more help over there. |
Beta Was this translation helpful? Give feedback.
-
I understand all the frustration. We're frustrated too. We know the root cause of this but it requires developing a complete replacement for a particular backend service we inherited. Right now we are working to meet commitments the United States has made to the World Meteorological Organization and its member states. So we are just having to limp along with that broken service for now and deal with it as needed. We are very close to the end of that WMO effort and then fixing this for good is a high priority in our 2021 development cycle. As noted previously, please report these issues to the NCO ops team. They do have ways to fix things temporarily. |
Beta Was this translation helpful? Give feedback.
-
Certainly understand the issues preventing a solution here. 2021 is 50% gone could we get an update even if it's just things are in the same place? Thank you |
Beta Was this translation helpful? Give feedback.
-
Thank you for the update!
…On Wed, Jun 30, 2021 at 9:20 AM Jared Holmberg ***@***.***> wrote:
The good news is we kicked off the sprint to refactor the service that
@StephenClouse <https://github.com/StephenClouse> mentioned which
directly impacts anything related to the netcdf grids. The bad news is that
it immediately got derailed supporting the team completing the QA process
for the other project he mentioned. On the plus side, the other effort puts
the entire program on better footing. We're not responsible for deploy, but
I estimate the other project should be released in a few months, and then
the refactor for the grid service a few months after that.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#329 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ATYUQZBE55UXZ2J6PBNN2LTTVM73RANCNFSM43JQCMEA>
.
|
Beta Was this translation helpful? Give feedback.
-
Is there an update on fixing this bug? As of 10/9/2021, I am still getting stale data. |
Beta Was this translation helpful? Give feedback.
-
Thanks for the update!
… On Oct 9, 2021, at 9:39 PM, Stephen Clouse ***@***.***> wrote:
Unfortunately we do not have an update. Our development schedule got hit with a nuke.
We are looking at some stopgap measures but no ETA.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications on the go with GitHub Mobile for iOS or Android.
|
Beta Was this translation helpful? Give feedback.
-
Is there a timeline for a fix on this yet? It would be great to have functioning again. |
Beta Was this translation helpful? Give feedback.
-
I'd reported it to nco.ops last year as INC0272004 . They've now merged in another (later one) from INC0322796 onto this Master ticket.
Maybe, the end is in sight? I hope so.. been a long slog. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
On some requests to gridpoint and/or zone forecasts like:
https://api.weather.gov/gridpoints/MTR/97,101/forecast
https://api.weather.gov/zones/forecast/CAZ513/forecast
very old/stale data will be returned from Akamai servers.
Currently, it seems to be happening with responses from 23.200.224.17, and 184.84.228.8 but intermittently.
A response from 23.200.224.17 for https://api.weather.gov/gridpoints/MTR/97,101/forecast returned headers of
and the data was 23:43:12 (h:m:s) old at the time.
A followon query to https://api.weather.gov/zones/forecast/CAZ513/forecast answered by 23.200.224.17 returned headers of
and actual data that had updated: at 127:43:35 (h:m:s) old.
It appears that not all the Akamai cache servers are regularly refreshing the gridpoint and/or forecast zone forecasts.
Beta Was this translation helpful? Give feedback.
All reactions