(I had a related post before but with wrong observation. I am now posting a new one instead of replying the old thread to avoid confusion)
SELECT ad_group_ad.ad.final_urls, ad_group_ad.ad.final_mobile_urls, ad_group_ad.ad.id,
ad_group_ad.ad.responsive_search_ad.descriptions,
ad_group_ad.ad.responsive_search_ad.headlines,
ad_group_ad.ad.responsive_search_ad.path1,
ad_group_ad.ad.responsive_search_ad.path2,
ad_group_ad.ad.tracking_url_template,
ad_group_ad.ad.url_custom_parameters, ad_group_ad.ad.final_url_suffix,
ad_group_ad.status, ad_group.id, ad_group.name, campaign.name
FROM ad_group_ad
WHERE campaign.id IN (some,campaign,ids)
AND ad_group_ad.ad.type = 'RESPONSIVE_SEARCH_AD'
AND campaign.status != 'REMOVED'
AND ad_group.status != 'REMOVED'
AND ad_group_ad.status != 'REMOVED'
ORDER BY ad_group_ad.ad.id ASC
I intentionally removed the long campaign id list.
We noticed that there have been various errors returned from similar queries:
So far this happen most frequently. It happens randomly after the timeout value for the query that succeeded in the past. There were also cases where the error was returned only within a few seconds after we sent the request. When this happened, retry would likely hit the same error.
This happened a few times. For one query with many retries, this error may be returned for one of the retries. But eventually the query failed with Deadline Exceeded.
- The service is currently unavailable
This is a temporary failure which can get recovered with retry
The issue with unstable queries started from 4/6. On average, it happens 1-2 times per hour. Before that, it never had such a high occurrence rate, only once per a few weeks, which we treated as API hiccup.
Is there a way that you can help track some of the calls if I can provide the exact timestamp when we sent the request and received the response with the query?
Thanks,
Gavin