From 8fef8fe7b00e6ae99f8788f8b57fbe2af40d30b1 Mon Sep 17 00:00:00 2001
From: Ben Kalman <kalman@users.noreply.github.com>
Date: Wed, 24 Aug 2016 09:47:09 -0700
Subject: [PATCH] Add perf bot script (#2393)

---
 tools/run_perf_builder.py | 84 +++++++++++++++++++++++++++++++++++++++
 tools/run_pr_builder.sh   |  2 +-
 2 files changed, 85 insertions(+), 1 deletion(-)
 create mode 100755 tools/run_perf_builder.py

diff --git a/tools/run_perf_builder.py b/tools/run_perf_builder.py
new file mode 100755
index 0000000000..6784fb6790
--- /dev/null
+++ b/tools/run_perf_builder.py
@@ -0,0 +1,84 @@
+#!/usr/bin/env python
+
+# Copyright 2016 Attic Labs, Inc. All rights reserved.
+# Licensed under the Apache License, version 2.0:
+# http://www.apache.org/licenses/LICENSE-2.0
+
+'''This script runs the Noms PR Jenkins jobs on http://jenkins-perf.noms.io:
+- http://jenkins-perf.noms.io/job/NomsMasterPerf
+- http://jenkins-perf.noms.io/job/NomsPRPerf
+'''
+
+import copy
+import os
+import os.path
+import re
+import subprocess
+
+# These are the 'go test' packages for perf testing.
+# Note that adding entires will actually run all tests in that package, not just the perf tests.
+PACKAGES = [
+    './samples/go/csv/csv-import',
+]
+
+# 'go test' timeout. Go's default is 10m, which ideally is long enough, but raise it a little.
+TIMEOUT = '15m'
+
+# Number of perf test repetitions. Choose 4 because:
+# - Any less than 3 is not enough sample size.
+# - Any more than 4 will take too long.
+# - Typically the first test will take longer, so choosing 4 gives you 3 non-first samples.
+PERF_REPEAT = '4'
+
+def main():
+    # Workspace is the root of the builder, e.g. "/var/lib/jenkins/workspace/NomsMasterPerf".
+    workspace = os.getenv('WORKSPACE')
+    assert workspace
+
+    # Directory where Go binaries have been installed.
+    go_bin = '/usr/local/go/bin'
+    assert os.path.exists(go_bin)
+
+    # Jenkins has arranged for the testdata directory to be in a shared workspace, as opposed to
+    # noms/../testdata like a normal checkout.
+    jenkins_home = os.getenv('JENKINS_HOME')
+    assert jenkins_home
+    testdata = os.path.join(jenkins_home, 'sharedspace/testdata/src/github.com/attic-labs/testdata')
+    assert os.path.exists(testdata)
+
+    # PRs have a "sha1" environment variable. This will actually look like "origin/pr/2393/merge",
+    # so extract the PR number to use as a prefix.
+    # For the master builder, just use the prefix "master".
+    pr_branch = os.getenv('sha1')
+    if pr_branch:
+        pr_pattern = re.compile(r'^origin/pr/(\d+)/merge$')
+        pr_groups = pr_pattern.match(pr_branch)
+        assert pr_groups
+        perf_prefix = 'pr_%s/' % (pr_groups.group(1),) # pr_2393/
+    else:
+        perf_prefix = 'master/'
+
+    # The database access token is given in a NOMS_ACCESS_TOKEN environment variable, in an attempt
+    # to hide it from the public Jenkins logs.
+    access_token = os.getenv('NOMS_ACCESS_TOKEN')
+    assert access_token
+
+    cmd = [os.path.join(go_bin, 'go'), 'test', '-timeout', TIMEOUT] + PACKAGES + [
+        '-perf', 'http://demo.noms.io/perf?access_token=%s' % (access_token,),
+        '-perf.repeat', PERF_REPEAT,
+        '-perf.prefix', perf_prefix,
+        '-perf.testdata', testdata]
+    cwd = os.path.join(workspace, 'src/github.com/attic-labs/noms')
+    env = copy.copy(os.environ)
+    env.update({
+        'GOPATH': workspace,
+        'PATH': '%s:%s' % (os.getenv('PATH'), go_bin),
+    })
+
+    proc = subprocess.Popen(cmd, cwd=cwd, env=env)
+    proc.wait()
+    assert proc.returncode == 0
+
+
+if __name__ == '__main__':
+    main()
diff --git a/tools/run_pr_builder.sh b/tools/run_pr_builder.sh
index 6e42fe1356..d040059f61 100644
--- a/tools/run_pr_builder.sh
+++ b/tools/run_pr_builder.sh
@@ -6,7 +6,7 @@
 
 # This script runs on the Noms PR Builder (http://jenkins.noms.io/job/NomsPRBuilder).
 
-set -e
+set -eux
 
 export GOPATH=${WORKSPACE}
 export PATH=${PATH}:/usr/local/go/bin:/var/lib/jenkins/node-v5.11.1-linux-x64/bin