McMahon - Last Good ECMWF was ( 2024031212 z ) and Last Should Be ( 2024031218 z )
Note: this was Checked at 2024-03-13 02:00 UTC.
It is possible that the Data was running Late and may now be available.
Note: this is an automated Post, check data to ensure it is correct!
Kindest Regards,
Tony
I’m already monitoring this. 4 out of the last 5 runs have completed successfully, but 20-30 minutes later than the running average completion times.
Out of the four ‘late’ runs three of them show the data becoming available 20+ minutes later than normal and one shows the data becoming available on time but taking 30+ minutes to download. Downloads usually take just a couple of minutes.
I’m suspecting problems with the ECMWF systems. A slow download could be caused by network issues on my server or the data centre network, but the test for data availability only requires a tiny bit of data transfer and if the network for my server was so bad at transferring a tiny amount of data, we’d almost certainly notice that in the performance of Discourse and other functions on the same server. For example, GFS processing (on the same server) is completing within 5 minutes of the running average times.
1 Like
The 00 and 12z ECMWF runs both seem to be running to minutes late. Everything else, including GFS seems to be running to time. The data availability time looks to be the cause - if the data isn’t available then I can’t download and process it.
I’ll take a look at the code to see if there’s a reason why the 00/12z runs wouldn’t see the data when it’s available.