mozilla / fission-monitoring-nightly

⚠️ This repository is archived per https://bugzilla.mozilla.org/show_bug.cgi?id=1738859

Code-base to generate monitoring dashboard for ongoing Fission Nightly "experiment".

Process

The analysis_etl.Rmd reads data in from the telemetry.main_nightly and telemetry.crash tables for performance, engagement, usage and stability metrics. 1st, it creates an temporary table to store client/day aggregate crash metrics. Next, it analyzes (e.g., aggregates and calculates confidence intervals) the probes relevant to the rollout, generating a 2nd ETL table moz-fx-data-shared-prod.analysis.fission_monitoring_dashboard_v1. The markdown results of the analysis ETL are available here.

The final step renders dashboard.Rmd, which pulls data from the previous analysis table and generates a plethora of Vega-lite plots. The rendered dashboard.html is uploaded to a GCS bucket, and then served by Protosaur.

Docker container

Building

docker build -t fission_monitoring_nightly .

Running

docker run -it -v PATH_TO_CREDENTIALS.json:/app/.credentials -e GOOGLE_APPLICATION_CREDENTIALS=/app/.credentials -e BQ_BILLING_PROJECT_ID=YOUR_BILLING_PROJECT -e BQ_INPUT_MAIN_TABLE=YOUR_BQ_INPUT_MAIN_TABLE -e BQ_INPUT_CRASH_TABLE=YOUR_BQ_INPUT_CRASH_TABLE -e BQ_OUTPUT_TABLE=YOUR_BQ_OUTPUT_TABLE fission_monitoring_nightly

Additional parameters:

• -e GS_BUCKET: Uploads dashboard.html to $GCS_BUCKET/dashboard/dashboard.html. In addition, uploads rendered ETL markdown to $GCS_BUCKET/etl/analysis_etl.html.
• -e DEBUG=true: debugging purposes (e.g., only process a couple probes to ensure ETL works as intended).
• -e MIN_BUILD_ID: Minimum app_build_id to process. Format: YYYYMMDD. This defaults to previous 10 days worth of builds, and consequently filters initial queries to previous 10 submission_date.