2017-10-13 ETL(s) prod failure
Date | |
Authors | Chris Mills (Unlicensed) |
Status | Complete |
Summary | As a result of ongoing development the main ETL pipeline and the revalidation ETL did not complete at their scheduled time. |
Impact | Stale/Incomplete data was being shown to users. |
Root Cause
I (Chris Mills (Unlicensed)) made some changes to the ETL pipeline as we begin to migrate over to using the Mirror in production. These changes were using incorrect syntax which caused the pipeline not to run. Further to this I moved the ETLs to run on the mirror prod rather than carry on using the consolidated DR and moved over too early.
Trigger
In an attempt to improve error handling in the pipeline I caused the overall issue.
Resolution
I rolled back my changes to the configuration files and re-ran the ETLs
Detection
I noticed the ETLs ran but did not error at 7:46am after making changes which I was unable to test the previous day.
Action Items
Action Item | Type | Owner | Issue |
---|---|---|---|
mitigate/prevent | |||
Timeline
Supporting Information
e.g. monitoring dashboards
, multiple selections available,
Slack: https://hee-nhs-tis.slack.com/
Jira issues: https://hee-tis.atlassian.net/issues/?filter=14213