As part of their work @Julia.glen needs access to the search logs that are stored in the analytics cluster. Specifically, the discovery.query_clicks_* and wmf_raw.cirrussearchrequestset tables in hive. Will need confirmation from legal about the contract end dates to set the revocation date appropriately.
requested group: analytics-privatedata-users
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Open | None | T212884 [EPIC] Improve Search Suggestions with NLP (Did You Mean / Glent) | |||
| Resolved | TJones | T212885 NLP contractor set up and access | |||
| Resolved | Dzahn | T214623 Analytics query access for search platform NLP contractor @Julia.glen | |||
| Duplicate | None | T214630 Volunteer / Collaborator NDA for @Julia.glen |
We also need to have an NDA on file that I am guessing will be signed after contract is, @TJones has the NDA signing taken place?
Hmmm. We may have glossed over that in the stand up. @Julia.glen said the contract was signed, but I don't know about a separate NDA. I don't have access to the legal documents, so I can't check, other than to ask legal, but I don't actually know who to ask.
For full details see https://wikitech.wikimedia.org/wiki/Production_shell_access#New_users
Will need from you (maybe more, sre to follow):
I support this request. @Julia.glen needs access to our search logs to design and build algorithms to improve the quality of our "did you mean" suggestions in search results. She needs to be able to mine the logs in large quantities to build out the various approaches (see T212884 for details on the various methods).
@Gehel Since we are all finally back I want to make sure this is progressing, it's very important for julia's work
@EBernhardson One reason that blocks progress is that no SSH public key has been pasted yet. The other is that due to Allhands nobody was / is on clinic duty. And access requests are usually handled by the on duty person.
ping @Julia.glen to paste ssh public keys in ticket (maybe a paste in phab that is best)
Sorry for the all hands delay. If the applicable contract is with Glenbrook Networks, Inc, it looks like that one was signed by Julia but not by Victoria before she departed. So let me follow up on getting it fully signed. I'm pretty sure the MSA has the appropriate non-disclosure language within it and will be sufficient. I'll check with the contracts attorney and get back on both points.
Change 488120 had a related patch set uploaded (by Dzahn; owner: Dzahn):
[operations/puppet@production] admins: create user with analytics-privatedata access for juliaglen
I talked on IRC with @EBernhardson and clarified the requested group is analytics-privatedata-users, updated ticket accordingly.
thanks for uploading a key @Julia.glen , i found you in LDAP and was able to upload the needed code change:
https://gerrit.wikimedia.org/r/c/operations/puppet/+/488120
This is pending review / approval now and i'll wait for Legal to confirm the NDA part.
Nuria said on Gerrit " Contract expires on May 31, 2019."
We should double-check which one is right.
As of today:
continuing until May 31, 2019."
Just confirming that the Master Services Agreement and Data Processing Agreement with Glenbrook are now fully signed. No separate NDA is needed, so you can proceed with access for julia.glen
We can probably use May 31, 2019 at the moment and extend it if the contract is extended. Worst case, we'll check the status on May 31 and update as needed.
Change 488120 merged by Dzahn:
[operations/puppet@production] admins: create user with analytics-privatedata access for juliaglen
puppet is creating her user on all the relevant servers right now. in max. 30 min it should all work
Change 512404 had a related patch set uploaded (by Dzahn; owner: Dzahn):
[operations/puppet@production] admins: remove expired contractor account of juliaglen (merge on May 31)
Change 512404 abandoned by Muehlenhoff:
admins: remove expired contractor account of juliaglen (merge on May 31)
Reason:
Merged 513954 instead