Closed jsrikish closed 4 weeks ago
Jayanthi to look into the size / NA
Jayanthi, Madhu, and Abdelhak have been working to identify the cause of the difference in fields between metadata in prod and uat. They have a plan forward, and aim to resolve tomorrow (Wed).
"units"="NA" was due to bytes not being a Unit option (wiki link) and best practices being to not later the size to an approximation. The full units/size is included in the metadata section that refers to the data itself (not URL to data)
Workflow was checked in multiple ways by Jayanthi, Abdelhak, Madhu.
Ingest subset of roughly 100-200 WV02_Pan_L1B granules into UAT to allow verification of metadata and metrics.
ACCOUNT: [csda-cumulus-uat-1686]
To find a date range that has roughly 100-200 granules, make use of AWS Athena by doing the following:
MSCK REPAIR TABLE maxar_transfer
dt
and the path pattern -- see notes below) until a count near the 100-200 range is obtained:This approach uses the count of
*BROWSE.jpg
files as a proxy for the number of granules in the date range.NOTE: the value of
dt
in the query above must be set to the most recent date of the S3 inventory (likely the current date, but the day prior to the current date should always work since the most recent 3 days are retained).NOTE: days like 2009/321 and 2009/339 (because the end date is exclusive in the rule definition) must be converted to
yyyy/MM/DD
for determining ISO 8601 datetime values for the start and end dates for the sample rule. This site provides conversion: https://dqydj.com/day-of-the-year-calculator/2009/11/17T00:00:00Z
2009/12/05T00:00:00Z
Acceptance criteria:
WV02_Pan_L1B___1_UAT.json
that spans a period of time that includes roughly 100-200 granules (using instructions above for finding appropriate values forstartDate
andendDate
)q
should be collection short name)