Changes to the WADP update procedure

r3naissanc3r

TNPer
-
-
Hi all,

We made a major overhaul to the way the WADP and associated data dumps and statistics (including the Endorse page, the WA activity reports, and all WADP censuses) are updated each day.

In the past, immediately after the end of the NS major update, the offsite server would use the NS API to pull data for each of the WA nations in TNP, one by one. This would be about 1300-1500 API calls, spread out to a period of about 45 minutes.

With the overhaul, all of this data is pulled from the NS nation dump, as soon as that becomes available, without the need for any API calls.

The changes have a few effects:
  1. In the past, even if NS successfully updated, the WADP could fail to update if NS became inaccessible during the one-hour period after the major update. Now, so long as the nation dump is generated successfully by NS, the WADP should also successfully update.
  2. In the past, there could be small glitches in the form of 1-2 missing nations or endorsement counts being off by 1-2 endorsements. Now everything should reflect the game snapshot immediately after the major update - any glitches will also be NS glitches.
  3. In the past, the endorse page did not update until around 3:30 am ET, and the WA reports until around 5 am ET. Now everything will update about 5 mins after the end of the major update.
  4. In the past, endorsements would continue to count towards the following day's WADP stats even if they came after the major update (and in particular, if they were exchanged between 1-2 am ET). Now WADP-related data is frozen as soon as TNP's update during the major update is over (currently, major update ends at around 1:10 am ET).
  5. In the past, the WA activity reports would record the SPDR and influence rank of every WA nation in the region. Now, they will only record influence rank.
  6. In the past, the WADP update cost us about 3000 API calls. Now it's essentially free.

Number (4) above should be important for those of you that actively compete for top spots in the WADP censuses. The well-defined cutoff, based on the regional update itself, should make rankings at the top few spots more predictable.

Overall, this change should make our WADP infrastructure a lot more efficient, robust, and fault tolerant. It should greatly reduce the occurrences of botched WADP updates like the one we experienced on Sunday. We do lose some information we were previously recording (see number (5) above), but all in all it should be a small loss.

This is a major overhaul - we had to rewrite a very large portion of the code running the WADP. Therefore, it's likely that some bugs may have been introduced. Please post here if you notice anything.
 
Last edited:
Back
Top