Filter out non-whitelisted experiment types and blocklisted experiment ids in moz_telemetry_s3

Author: Sunah Suh

moz_telemetry/CMakeLists.txt

@@ -3,7 +3,7 @@
 # file, You can obtain one at http://mozilla.org/MPL/2.0/.
 
 cmake_minimum_required(VERSION 3.0)
-project(moz-telemetry VERSION 1.2.14 LANGUAGES C)
+project(moz-telemetry VERSION 1.2.15 LANGUAGES C)
 set(CPACK_PACKAGE_DESCRIPTION_SUMMARY "Mozilla Firefox Telemetry Data Processing")
 set(CPACK_DEBIAN_PACKAGE_DEPENDS "${PACKAGE_PREFIX}-moz-ingest (>= 0.0.1), ${PACKAGE_PREFIX}-lsb (>= 1.1.0), ${PACKAGE_PREFIX}-circular-buffer (>= 1.0.2), ${PACKAGE_PREFIX}-heka (>= 1.1.9), ${PACKAGE_PREFIX}-elasticsearch (>= 1.0.3), ${PACKAGE_PREFIX}-rjson (>= 1.1.0), ${PACKAGE_PREFIX}-lfs (>= 1.6.4)")
 string(REGEX REPLACE "[()]" "" CPACK_RPM_PACKAGE_REQUIRES ${CPACK_DEBIAN_PACKAGE_DEPENDS})

moz_telemetry/sandboxes/heka/output/moz_telemetry_s3.lua

@@ -49,6 +49,10 @@ preserve_data       = not flush_on_shutdown -- should always be the inverse of f
 
 -- Specify an optional module to encode incoming messages via its encode function.
 -- encoder_module = "encoders.heka.framed_protobuf" -- default
+
+-- Specifies experiment types whitelist, and experiment id blocklist
+experiment_types = {["normandy-preference-"] = true} -- optional
+experiment_blocklist = {["pref-flip-screenshots-release-1369150"] = true} -- optional
 ```
 --]]
 
@@ -85,6 +89,10 @@ if experiment_dimensions then
     experiment_batch_dir  = read_config("experiment_batch_dir") or error("experiment_batch_dir must be specified")
     experiment_dimensions = mts3.validate_dimensions(experiment_dimensions)
 end
+local experiment_types = read_config("experiment_types") or {}
+assert(type(experiment_types) == "table", "experiment_types must be a table")
+local experiment_blocklist = read_config("experiment_blocklist") or {}
+assert(type(experiment_blocklist) == "table", "experiment_blocklist must be a table")
 
 
 local function get_fqfn(dir, path)
@@ -190,6 +198,12 @@ local function process_standard_dimensions(data)
 end
 
 
+local function does_experiment_qualify(id, branch, experimentType)
+    return (not experimentType or experiment_types[experimentType])
+        and not experiment_blocklist[id]
+end
+
+
 local function process_experiment_dimensions(data, experiments)
     local ok, experiments = pcall(cjson.decode, experiments)
     if not ok then return end
@@ -198,11 +212,13 @@ local function process_experiment_dimensions(data, experiments)
     for id, exp in pairs(experiments) do
         vars.experimentId = id
         vars.experimentBranch = exp.branch
-        local dims = {}
-        for i,d in ipairs(experiment_dimensions) do
-            dims[i] = mts3.read_dimension(d, vars)
+        if does_experiment_qualify(id, exp.branch, exp.type) then
+            local dims = {}
+            for i,d in ipairs(experiment_dimensions) do
+                dims[i] = mts3.read_dimension(d, vars)
+            end
+            output_dimension(experiment_batch_dir, dims, data)
         end
-        output_dimension(experiment_batch_dir, dims, data)
     end
 end
 

moz_telemetry/tests/hindsight/run/input/moz_telemetry_s3.lua

@@ -12,7 +12,8 @@ require "string"
 local inputs = {
     {Timestamp = 0, Type = "moz_telemetry_s3", Fields = { docType = "main"}},
     {Timestamp = 0, Type = "moz_telemetry_s3", Fields = { docType = "main", ["environment.experiments"] = '{"foo":{"branch":"123"}}'}},
-    {Timestamp = 0, Type = "moz_telemetry_s3", Fields = { docType = "main", ["environment.experiments"] = '{"foo":{"branch":"123"}, "bar":{"branch":"456"}}'}}
+    {Timestamp = 0, Type = "moz_telemetry_s3", Fields = { docType = "main", ["environment.experiments"] = '{"foo":{"branch":"123"}, "bar":{"branch":"456"}}'}},
+    {Timestamp = 0, Type = "moz_telemetry_s3", Fields = { docType = "main", ["environment.experiments"] = '{"foo":{"branch":"123", "type":"normandy-preference-"}, "bar":{"branch":"456", "type":"another-type"}, "pref-flip-screenshots-release-1369150":{"branch":"789"}}'}}
 }
 
 
@@ -46,9 +47,10 @@ function process_message()
         local ed = eh:read("*a")
         eh:close()
 
-        if od:match("main\n") and ed:match("main%+foo%+123") and ed:match("main%+bar%+456") then
-            check_message_count("output/moz_telemetry_s3/main", 3)
-            check_message_count("output/moz_experiments_s3/main+foo+123", 2)
+        if od:match("main\n") and ed:match("main%+foo%+123") and ed:match("main%+bar%+456")
+            and not ed:match("main%+pref-flip-screenshots-release-1369150%+789") then
+            check_message_count("output/moz_telemetry_s3/main", 4)
+            check_message_count("output/moz_experiments_s3/main+foo+123", 3)
             check_message_count("output/moz_experiments_s3/main+bar+456", 1)
             return 0
         end

moz_telemetry/tests/hindsight/run/output/moz_telemetry_s3.cfg

@@ -8,6 +8,8 @@ dimension_file  = "moz_telemetry_s3.json"
 
 -- https://bugzilla.mozilla.org/show_bug.cgi?id=1353110
 experiment_dimension_file  = "moz_telemetry_s3_experiments.json" -- optional
+experiment_types = {["normandy-preference-"] = true} -- optional
+experiment_blocklist = {["pref-flip-screenshots-release-1369150"] = true} -- optional
 
 -- directory location to store the intermediate output files
 batch_dir               = "output/moz_telemetry_s3"