Skip to content

Latest commit

 

History

History
31 lines (26 loc) · 1.63 KB

experiment.md

File metadata and controls

31 lines (26 loc) · 1.63 KB
Raw

Experimental feature for Aggie

To allow for controlled experiments and evaluation of Aggie, we have developed a content service that will replay the reports found in a json file.

Reports will be "replayed" with the same cadence than in the original "recording". To set up and save a recording

  1. Start Aggie normally with the sources you want to record.
  2. Set fetching off when you want to finish your recording.
  3. Dump the sources mongodump --db aggie --collection sources as we need to use the same mongoDB's _id.
  4. Export the reports you want to replay to a json file for the time range you want e.g: mongoexport --db aggie -c reports -q '{storedAt: { "$gt": {"$date" : "2016-11-14T03:54:26.130Z"}, "$lt": {"$date" : "2016-11-14T03:54:30.130Z"}}}' --sort '{storedAt: 1}' --sort '{storedAt: 1}' > reports.json
  5. Export the previous reports: mongoexport --db aggie -c reports -q '{storedAt: { "$lte": {"$date" : "2016-11-14T03:54:26.130Z"}}}' --sort '{storedAt: 1}' > previous_reports.json

To play a recording:

  1. Set fetching off
  2. Stop Aggie.
  3. Drop sources, incidents, reports. Drop other collections if not needed in your experiment.
  4. Load previous reports if needed: mongoimport --db aggie --collection reports --file previous_reports.json
  5. Set the experimental variables of secrets.json: experiment to true and point experimentFile variable to where the json file is, relative to the Aggie folder.
  6. Restore the sources dump to the database: mongorestore --db aggie --collection sources dump/aggie/sources.bson
  7. Start Aggie.
  8. Set fetching on.
  9. We are now replaying reports.json.