Third-party modules symbolication - GCP restart
Categories
(Data Platform and Tools :: General, task, P3)
Tracking
(Not tracked)
People
(Reporter: akomar, Unassigned)
Details
We need a symbolicated version of Third Party Modules ping table.
This is the next iteration of untrusted-modules ping, which we were attempting to symbolicate here.
We can either repurpose the old job to run with BigQuery or rewrite this to Beam.
|
Reporter | |
Comment 1•6 years ago
|
||
The old Spark application lives here: https://github.com/mozilla/telemetry-batch-view/blob/untrusted-modules-symbolication/src/main/scala/com/mozilla/telemetry/views/untrustedmodules/UntrustedModulesView.scala
We were scheduling it from Databricks: https://dbc-caf9527b-e073.cloud.databricks.com/#job/1045
Data volume we have for Third Party Modules ping is fairly low, so we should be able to at least start testing this on Databricks, leveraging recently added support for writing dataframes in Spark-BigQuery connector.
|
|
Reporter | |
Comment 2•6 years ago
|
||
Toshihito: I imagine the best way for analyzing symbolicated stacks would be to have a table with the same fields you now have in third_party_modules + a column for stacktraces?
|
||
Comment 3•6 years ago
|
||
Currently we have Iodide report to view the Third Party Modules ping. If we can have a table or view integrated with symbolicated stacks, which can be accessed from the report, it's really nice.
|
|
Reporter | |
Comment 4•6 years ago
|
||
This is on hold for now. Toshihito managed to symbolicate stacks on demand directly from the notebook, we'll see if that's enough for his analyses.
|
||
Comment 5•4 years ago
|
||
Hasn't been touched in 3 years so closing as inactive. Please re-open if you have recent updates to consider doing this work.
|
Assignee | |
Updated•3 years ago
|
Description
•