1594111 - Mission Control v1 should use bigquery instead of athena

Status: RESOLVED FIXED

(Cloud Services :: Mission Control, task)

Description

[William Lachance (:wlach)]

So that we can decomission error_aggregates, we should rework mission control to get data from a bigquery view generated from the live ping tables (bug 1594108)

Comment 1

[William Lachance (:wlach)]

Ben Wu is leading this, assigning to him. CC'ing a few people from relman to keep them updated on this.

FYI for relman: There will be a few user-visible changes once this is deployed to production (probably in a few weeks):

• Data will probably be updated a little less frequently (probably every few hours)
• We were double counting some crash / usage information due to a bug in the way we were handling experiments which Ben discovered when doing this work. We are pretty sure this didn't materially affect results (since the double counts would effectively cancel each other out), but it's something to bear in mind. In any case, the new bigquery-based dataset Ben is working on will not have this problem.
• We will no longer be updating non-crash measures in the mission control dashboard (e.g. slow_script_count). These were never used/discussed in the channel meetings.

If you have any questions feel free to ask either here or in the #stability slack channel.

Comment 2

[William Lachance (:wlach)]

We can mark this as fixed, missioncontrol.tmo is now using bigquery with no major issues